6 'P's in AI Pods (AI6P): People & Places

📜 AISW #105: Joan Kinyua, Kenya-based digital rights advocate

Karen Smiley — Thu, 09 Apr 2026 10:07:03 GMT

Introduction - Joan Kinyua

This week’s AI, Software, & Wetware features a written interview with Joan Kinyua, a 🇰🇪 Kenya-based digital rights advocate. Joan is the founding President of the Data Labelers Association, which was recently profiled by 404 Media [“AI Is African Intelligence': The Workers Who Train AI Are Fighting Back”, by Jason Koebler, Mar 12, 2026]. We discuss:

Her experience as a data worker and how that prompted her to found the Data Labelers Association
How she does & doesn’t use AI tools to help with her advocacy work
Examples of misalignment between data work tasks and local context or Kenyan ways of understanding situations
How an AI tool caused her to lose an account, with only a facade of an appeal process
Ways people can work with the Data Labelers Association to support digital rights

and more. Check it out, and let us know what you think!

Leave a comment

This post is part of our AI6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and works. (This is a written interview; read-aloud is available in Substack. If it doesn’t fit in your email client, click HERE to read the whole post online.)

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Joan Kinyua

Karen: I am delighted to welcome Joan Kinyua from Kenya as my guest today on “AI, Software, and Wetware”! Joan, thank you so much for joining me on this interview. Please tell us about yourself, who you are, and what you do.

Joan: Thank you so much for having me! I’m Joan Kinyua, from Nairobi, Kenya. I’m a digital rights and labor advocate and the founding President of the Data Labelers Association, a grassroots, worker-led organization that represents data workers who label datasets for AI systems. My work focuses on advocating for ethical labor practices, fair pay, transparency, and mental well-being for people working in the AI value chain.

I started my journey as a data worker myself, so my advocacy comes from lived experience. I’ve seen firsthand the challenges, precarious working conditions, and lack of recognition that data workers face. Through the Association, we aim to make this invisible workforce visible, valued, and protected, ensuring that the human labor behind AI is recognized as central to technological development, not peripheral.

Karen: What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

Joan: My experience with AI, machine learning, and analytics is primarily from the perspective of a data worker and labor advocate, rather than as a developer or engineer. I have extensive hands-on experience in data annotation, labeling datasets that are used to train AI models for social media, autonomous vehicles, and other applications. Through this work, I gained a deep understanding of the types of data AI systems rely on, how datasets are structured, and the labor required to make AI function effectively.

In addition, I have studied and engaged with AI and ML concepts to advocate for ethical practices and fair labor standards. This includes analyzing the impact of AI on labor markets, understanding biases in algorithms, and exploring how governance frameworks can protect workers in the AI ecosystem. While I’m not building AI models myself, my professional and research experience allows me to bridge the technical and human sides of AI, highlighting the critical role of the workforce that powers these systems.

Karen: That’s critically important work. Are AI tools ever of help to you in doing that work? I hear from a lot of entrepreneurs that they find AI tools can save them time on the many tasks of running a business or organization.

Joan: Yes, to some extent they are helpful. We primarily use AI tools for grammatical structuring and refining sentence flow, especially when preparing formal communications, reports, or policy documents. They are also useful when we need to adapt language for specific audiences, ensuring the tone and framing are appropriate for different stakeholders, whether policymakers, workers, or international partners.

At an organizational level, AI tools can also support basic administrative functions such as bookkeeping and record organization, helping streamline internal processes. However, their use remains complementary, the core of our work, particularly around advocacy and worker engagement, is still deeply human-centered.

Karen: I’m also curious if you ever use AI tools in your personal or community life? (AI is hard to avoid in daily life - see “But I don’t use AI”: 8 Sets of Examples of Everyday AI, Everywhere)

Joan: Yes, I do use AI tools in my personal and community life, particularly as a way to make complex information more accessible. For example, when I encounter concepts or discussions that are difficult to understand, I use AI to break them down into simpler, clearer explanations.

This is especially helpful when dealing with heavy legal or litigation language, which can often be dense and difficult to interpret. AI tools help translate that into more understandable terms, making it easier to engage with and share within community contexts.

That said, it’s mainly a support tool — it helps with clarity and accessibility, but the interpretation and application still come from lived experience and collective understanding.

Karen: Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

Joan: Yes! I actually worked on a large language model (LLM) project, where my task was to prompt the model with a question, review the response, and select the answer that best fit the prompt. The project was in Kiswahili, which made it particularly interesting, and in many ways, I can see a direct parallel with tools like ChatGPT.

What struck me was how the model seemed to extract so much from me as a human worker. It relied heavily on my judgment and contextual understanding to evaluate responses correctly. On the positive side, it was impressive how the LLM could generate coherent answers and handle linguistic nuance, especially in Kiswahili, which is less commonly represented in AI datasets. On the other hand, challenges included ambiguities in the responses, occasional irrelevance, and the model’s tendency to overgeneralize, which meant I had to pay careful attention to quality and correctness. Overall, it was a fascinating experience that highlighted the essential role human judgment plays in AI systems, especially in non-English contexts.

Karen: Aside from the challenges you noted with the quality of the answers, were you ever asked to handle any questions that just didn’t align with typical Kenyan values or ways of thinking? For instance, if the question had a built-in assumption that meant none of the answers would really be ‘right’ for a Kenyan person who would be using the tool in Kiswahili, even if the answers were syntactically correct.

Joan: Yes, I would respond to that in a general way.

At times, the tasks and instructions we were given came with predefined assumptions that didn’t necessarily align with local context or Kenyan ways of understanding situations. As data workers, we were expected to strictly follow the instructions, even when they didn’t fully reflect the reality on the ground.

For example, content such as photos or videos of protests could be framed in a way that emphasizes instability or insecurity, suggesting that a country is unsafe. However, from a local perspective, these protests are often rooted in legitimate grievances and injustices that people are actively addressing. That nuance is not always captured in the task framing.

So even if the outputs were technically or syntactically correct, they could still misrepresent context, reinforce bias, or miss important cultural and social realities. This highlights a broader issue in AI systems, the gap between standardized instructions and lived, localized understanding.

Karen: Those are great examples - thank you, Joan.
If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Joan: Yes, there are areas where I intentionally avoid using AI tools.

I generally avoid relying on AI for things that require deep human context, emotional nuance, or professional judgment. For example, matters involving feelings, personal experiences, or relationship decisions are not something I would trust AI with, because these require empathy, lived experience, and an understanding of context that a machine cannot fully grasp.

Similarly, I would not rely on AI for medical advice or prescriptions, where accuracy, accountability, and professional expertise are critical.

In these cases, I see AI as just an interface or support tool, not a substitute for human judgment, care, or expertise.

Karen: A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Joan: I am of the firm opinion that if data or content does not belong to you, it should never be used without explicit permission. AI and ML models are not inherently intelligent, they are entirely dependent on human labor, input, and creativity. Yet too often, companies extract this work without consent, recognition, or compensation, creating vast asymmetries of power where humans provide the intelligence and the AI takes the credit.

Ethical AI companies must adhere to the “3Cs Rule”: Consent, Credit, and Compensation. Users and workers whose data powers AI systems should have full agency over whether their data is used, be acknowledged for their contributions, and receive fair remuneration. Without these safeguards, AI development perpetuates exploitation and erodes accountability, ignoring the very people whose labor and knowledge make these technologies functional. Respecting these principles is essential for building AI systems that are genuinely responsible, sustainable, and equitable.

Karen: As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Joan: No, in my experience, AI tool providers are generally not transparent about where their training data comes from or whether the original creators have consented to its use. For example, many large language models claim to use “publicly available” data, but they rarely clarify which sources were included, whether creators were informed, or if any compensation was provided.

This lack of transparency is concerning because it obscures the human labor and creativity that underpin these systems, effectively allowing companies to benefit from the work of others without accountability. For AI to be ethical, users and creators must know where the data comes from, that their consent has been given, and that they are recognized and compensated for their contributions.

Karen: (And that the data was annotated and processed by workers who are treated ethically ;) If you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

Joan: Honestly, most of the time, I have no idea where the data is coming from, nor do I fully understand the model I’m contributing to. To make it even more surreal, the project names themselves are almost always cryptic, often just an animal or a random jumble of letters, so it’s impossible to make sense of what they’re actually building. We’re handed tasks, complete them, and move on, without context or insight into the final product. It really shows how opaque AI development is, and how the human labor powering these systems is treated as invisible, even though, without us, these models wouldn’t function at all.

Karen: As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Joan: Yes, it’s very likely that people’s personal data or content has been used by AI systems, though it’s often hard to pinpoint exactly when or how. Many of the tasks we work on are piecemeal and fragmented, which makes it difficult to understand the bigger picture of what we’re contributing to. Because each task is so small and isolated, it can feel almost insignificant, and we rarely have a clear sense of what information we are being exposed to or how it is being used. This fragmentation obscures the true scale of human labor behind AI and the potential privacy implications for the people whose data powers these systems.

Karen: Has your own data ever been used in an AI tool that you’re aware of?

Joan: I’m not entirely sure.

Like many people, I use platforms such as LinkedIn, email services, and other digital tools that may incorporate AI in different ways. However, it’s often not clear how or when personal data is being used, or to what extent it contributes to training or improving these systems.

So while it’s possible that my data has been used in some capacity, there isn’t enough transparency for me to say that with certainty.

Karen: Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML?

Joan: No, I do not know of any company that explicitly informed me that the data or content I provided would be used to train AI or ML systems. In most cases, there is no transparency or disclosure, and workers or users are left unaware of how their input is being leveraged.

Karen: Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Joan: No, I have not experienced issues like privacy breaches or phishing directly. However, I have had my account terminated arbitrarily, with the company stating I was in breach of their policies. This decision was made by an AI system, which highlights a troubling reality: the very systems we are helping to train and teach are increasingly making decisions that affect our own livelihoods. It’s a stark reminder of how human labor powers these models, yet the models themselves can exert control over the humans contributing to them.

Karen: That sucks that your account was terminated. Did you have any recourse when it happened? Was there any human review or appeal process that you could use to regain it?

Joan: No, there wasn’t any meaningful recourse. In fact, the appeal process often feels like a façade: you submit your request, but there is little to no response, and no clear indication that a human is actually reviewing your case.

Once the decision is made, it’s essentially final. There’s no transparency around why it happened, no opportunity to properly challenge it, and no real pathway to regain access. This lack of accountability can be incredibly frustrating, especially when your work and livelihood depend on these platforms.

Karen: Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Joan: The most important thing AI companies must do to earn and maintain trust is to recognize the human labor, creativity, and knowledge that power their systems, and act transparently and responsibly. Too often, AI models operate as black boxes, built on data, artwork, music, and creative content extracted without consent or compensation, while the people who make these systems possible remain invisible.

Ethical AI companies must center humans in their governance, ensuring that workers, artists, and contributors have a voice in how systems are designed, deployed, and monitored. Transparency is not just about sharing technical details, it’s about acknowledging the labor behind the AI, explaining decisions that affect lives, and being accountable for harm. Until AI companies take these steps, public trust will remain fragile, and the human foundation of AI will continue to be undervalued and exploited.

Karen: Thank you so much for making time for this interview, Joan! Is there anything else you’d like to share with our audience?

Joan: I’d encourage anyone interested in learning more about our work to visit our website at www.datalabelers.org.

We are always open to partnerships and collaborations with organizations, researchers, policymakers, and industry actors who are committed to advancing ethical AI and data worker rights. Please feel free to reach out through the website to connect with us.

Interview References and Links

Joan Kinyua on LinkedIn

Joan Kinyua on Bluesky

Data Labelers Association website and on LinkedIn

Leave a comment

About this interview series and newsletter

This post is part of our AI6P interview series on “AI, Software, and Wetware”. It showcases how real people around the world are using their wetware (brains and human intelligence) with AI-based software tools, or are being affected by AI.

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

We want to hear from a diverse pool of people worldwide in a variety of roles. (No technical experience with AI is required.) If you’re interested in being a featured interview guest, anonymous or with credit, please check out our guest FAQ and get in touch!

6 'P's in AI Pods (AI6P) is a 100% human-authored, 100% reader-supported publication. (No ads, no affiliate links, no paywalls on new posts). All new posts are FREE to read and listen to. To automatically receive new AI6P posts and support our work, consider becoming a subscriber:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it. All works, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)

📜 AISW #100: Karo Zieminski, Denmark-based product leader and builder

Karen Smiley — Thu, 05 Mar 2026 11:06:46 GMT

Introduction -

This edition of AI, Software, and Wetware features a written interview with Karolina Zieminski, a 🇩🇰 Denmark-based product leader and builder. Karo is the founder and author of the Substack bestseller Product with Attitude and the creator of StackShelf.app.

This post celebrates Karo’s 1-year anniversary on Substack, my 2-year anniversary on Substack, and the 100th interview in the AI, Software, & Wetware series. We discuss:

Studying AI obsessively and learning about AI by doing, not in the abstract
The prompt she uses to get review feedback on drafts of her articles from three perspectives
How she and her husband are guiding their kids on using AI tools responsibly
Creating a new legal framework to handle machine vision data sourcing responsibly
Why she started her 2026 “Builders With Attitude” initiative

and more. Check it out, and let us know what you think!

Leave a comment

This post is part of our AI6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and works. (This is a written interview; read-aloud is available in Substack. If it doesn’t fit in your email client, click HERE to read the whole post online.)

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

Karen: I am delighted to welcome Karolina Zieminski from Denmark as my guest on AI, Software, and Wetware! Karo, thank you so much for joining me on this interview. Please tell us about yourself, who you are, and what you do.

Karo: Thank you for having me! It feels a bit surreal to be on the other side of this, especially with you. You were one of the first people I met on Substack, one of my earliest subscriptions, and honestly a big part of how I figured out what this place is about.

By trade, I’m an AI product manager. By inclination, I’m a builder and a Substack writer. I spend most of my time where AI systems, product thinking, and human judgment bump into each other. Both in corporate environments and in public. That’s what I end up writing about in Product with Attitude. Sometimes very explicitly, and sometimes by sneaking product thinking into things like vibe coding guides.

I’m also a wife, a mom, and a dog-mom.

Karen: Karo, I remember when you first started on Substack in February 2025. Your writing and your drawings stood out immediately. I could tell you were going to succeed at leading and building a newsletter and community here, and you have! In less than a year, Product with Attitude became a Substack Bestseller with hundreds of paid subscribers, and I love seeing you on the ‘Rising’ leaderboard for Technology. So well deserved! 🙂

Karen: What is your level of experience with AI, ML, and analytics? I’d like to hear about how you have used it professionally or personally, or studied the technology, or built tools using the technology.

Karo: I’ve worked with AI professionally for over eight years, across product strategy for computer vision and physical AI, and through daily collaboration with ML engineers and researchers. I’ve also spent a lot of time implementing AI tools into real team processes. The unglamorous part where things either work or don’t.

I use AI every day. It would honestly be easier to list the things I don’t use it for than the things I do. Every part of my workflow, from research and brainstorming to productivity and coding, is supported by AI. I even drive a car with self-driving capabilities, so AI shows up in my life well beyond work.

That’s also why responsible tech matters so much to me. I’m not just commenting on it — I’m one of the people building it, and I want to do it right.

And yes, I study the technology, almost obsessively at this point. I don’t think any of us are immune to the career shifts AI is creating, not really. This is just how I deal with that reality.

Karen: Do you have any tips you can share with people who are just getting started on how they can learn and study AI? What worked for you, and how do you learn nowadays?

Karo: Love this question.

I learned more about AI by doing than by studying it in the abstract. My biggest tip for beginners: pick one real task you already care about, commit to using one AI tool on it every day for a week, and see what works. Then ask your favourite AI to help you turn your results into a repeatable workflow.

That will be your first automation. And once you see what’s possible, new ideas will follow, I promise. I’d also encourage everyone to join communities like Product with Attitude and document your journey, so others can support you along the way. Plus, it’s more fun this way.

Karen: Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
(Reference for this question: “But I don’t use AI”: 8 Sets of Examples of Everyday AI, Everywhere)

Karo: There are a lot of examples. One very concrete one is my daily use of ChatGPT as a thinking partner. I’ve trained it to push back on my assumptions and understand my thinking patterns better. To ask questions I was hoping to avoid. It’s not always comfortable and it definitely slows me down, but that’s usually where the breakthroughs happen - right around the moment I realize I was wrong.

Karen: Can you share an example of something specific you asked ChatGPT about, and what assumptions it helped you understand that you were making?

Karo: Not from the top of my head, but I have a prompt I use after I draft my articles:

Read this text as someone wearing three different hats: an expert, a beginner, and a skeptic. For each perspective, tell me what’s unclear, what’s missing, and what feels wrong.

It’s a simple trick, but it catches blind spots I’d never find on my own.

Karen: Do you have any examples you can share of a time you used an AI tool and it did NOT work well for your needs? For instance, has your self-driving car ever made any mistakes? Or has a chat tool ever given you bad advice on medical or veterinary questions?

Karo: Oh yes, so many times! It broke my code, deleted my files, gave me hallucinated answers. That’s why developing critical AI literacy matters so much. I wrote more about this in a recent article.

Product with Attitude

You Can't Critique AI You Haven't Built With

2 months ago · 144 likes · 42 comments · Karo (Product with Attitude)

(4) Karen: If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Karo: Yes, I have.

If a product makes it onto my Absurd Product Decision list, that’s a hard no. I won’t use it, recommend it, or accept a sponsorship request.

Product with Attitude

2025’s Most Absurd Product Decisions

4 months ago · 109 likes · 39 comments · Karo (Product with Attitude)

The same goes for social platforms with a reputation for exploitative data practices. I won’t join it, no matter how popular it is.

And I also stop using tools the moment I notice dark patterns during testing. If the ethical baseline isn’t there, I’m not interested.

Karen: So that’s a good view of which tools you choose to use or not use. How about use cases? 🙂 Is there any purpose that you would never use any AI tool for? (For instance, writing a personal email to a family member or friend, or creating music or images)

Karo: I’d never use AI for anything I wouldn’t do in real life. It’s that simple. If I wouldn’t plagiarize someone’s work, generate a fake review, impersonate someone, or manipulate people’s emotions. AI is just a tool, it doesn’t change the ethics.

Karen: You mentioned being a mom. How are you guiding your kids on using and learning AI? Is AI being used in their schools?

Karo: This is one of the most complex questions we face as parents. First of all, I’d never claim I have all the answers. I don’t think anyone does, no matter how confidently they post about it online. We simply can’t predict what AI will look like in 20 years, which means we can’t predict the full impact on the kids growing up with it right now. That’s a humbling thing to sit with.

Our approach at home is controlled exposure. My kids are 14 and 11, and they’re naturally AI-curious. My husband and I both work in tech, so they overhear the conversations, they see us experimenting, they ask questions.

When one of them wants to learn something, I’ll show them how AI can help with that, but I also explain the risks. Every single time. There’s a lot of repeating myself heheh. But I know from experience that understanding doesn’t come from one perfect conversation, it comes from hearing the same message in different contexts, over and over, until it clicks.

That’s the best I’ve got, and I think being honest about that uncertainty is more useful than pretending we have a parenting playbook for something none of us have seen before.

Karen: A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Karo: This is a very real concern. From my professional work, I know these practices are far more common across AI systems than most people realize, and they tend to live in broad grey zones.

Getting valuable data with full, informed consent is HARD. Enabling others to do it right is even harder.

It requires robust legal frameworks and a clear exchange of value for the people sharing their data. That’s exactly what we’ve been building in computer vision, and it was really, really hard. But if you want to build things the right way, you do hard things.

Karen: It’s great to hear that you helped to drive respect for data dignity in your work!
As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Karo: I honestly doubt that any of the tools I use have 100% provable data lineage, especially when it comes to data collected in the past, before regulations really tightened. BUT! I do think companies are getting better at understanding that transparency = trust. And more of them are starting to act on that.

Karen: As you pointed out, it’s often hard to know where companies get their data, because they’re often not transparent about it. As an AI tool user, does how a company acquires their data, and whether they respect these principles, factor into how you choose a tool? Or, if a company was transparent or if their process was proven in a way that we could actually know, would that factor into your tool choices in some way? Those decisions are always tradeoffs.

Karo: Yes, absolutely. No company is perfect on this front, so for me it comes down to transparency. If a company is open about how they source and handle data, I’ll invest in their tools. If they’re not, I move on. That’s the most practical stance I can take right now.

Karen: Can you share about where the data came from and how it was obtained when you’ve built an AI-based tool or system?

Karo: As I mentioned, it was really hard. To give some context, we had to build an entirely new legal framework, because there wasn’t one to rely on.

That process alone took over two years and involved multiple legal firms, ethical and regulatory bodies, customers, engineers. We wanted to go beyond just satisfying legal requirements and make sure that the value exchange was fair.

Karen: I love hearing about companies that are trying to handle data sourcing responsibly! Can you elaborate a bit on what kind of data was involved, and how the value exchange was eventually handled?

Karo: Yes, I could talk about this for hours! We’ve built a library of responsible data for computer vision.

By responsible, I mean the data is collected with informed consent, meets the highest GDPR standards, complies with human rights principles, the EU AI Act, and local regulations, is fully anonymized, and is never directly accessible to humans during model training. And all data is collected in exchange for access to better models.

Everyone wins:
- People who share data do so because they want to benefit from more accurate models. And they can do it securely, knowing their data is protected.
- And developers get access to data they wouldn’t be able to obtain otherwise, but they can never download it, and no private or sensitive information is ever exposed to them.

Nothing like this existed for computer vision before, not at this scale. And building it was really, really hard.

Karen: As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Karo: Yes, almost certainly. Like most people, my data has flowed through social media platforms, newsletter platforms, and analytics tools: signup information, posts, photos, probably behavioral data as well.

None of this is unusual, which is exactly the problem. Much of our data participation in AI systems happens quietly, almost by default, and through everyday interactions rather than dramatic moments of consent.

Karen: I agree - concepts of informed consent are problematic in the way terms and conditions are presented, and we often don’t have any real choice. But as a US resident, I tend to assume that people in Denmark have somewhat better protection. Do you find that regulations like GDPR and the Digital Markets Act are having an impact on making companies respect your personal data? Or are they not as helpful as people outside the EU might assume they are?

Karo: Absolutely. In Denmark, GDPR, the EU AI Act and other regulations are a part of training for everyone and we have corporate mechanisms to act if we see something wrong. It’s business as usual, a constraint everyone respects. To the point that when someone in our neighbourhood photographed a person stealing a bike and shared it in a local Facebook group rather than giving it to the police, they faced a backlash.

Karen: Are there any tips you can share with people on steps they could take to protect their data?

Karo: Yes, keep it close to you. Don’t share anything you wouldn’t want strangers to know about you, ever.

Karen: Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

Karo: Meta and Linkedin are good examples, where changes to terms around AI training were announced long after people had already shared years of content.

Karen: Yeah, the LinkedIn opt-out and Meta opt-outs in summer 2024 are great examples of this. Opt-out requests from European citizens had to be honored while ours did not. I ended up spending hours deleting my Facebook posts and photos before the deadline.
Are there any special steps you are taking to protect the privacy of your kids’ data? Any advice that you could share with other parents?

Karo: I’m in the ‘no social media for kids and teens’ camp, heheh. Not by the kids, not by the parents. Not until their prefrontal cortex is developed at 25 and they can make this decision for themselves. Luckily, where we live everyone agrees, so my kids aren’t the odd ones out.

Karen: Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Karo: Luckily, no. I haven’t seen any concrete negative impacts so far, although I’m aware that these issues often surface indirectly or over time.

Karen: You are lucky and I hope that holds up! Most Americans I know have been caught up in at least one data breach, and sometimes many. And we get lots of junk calls and spam and phishing emails from companies that bought our info from a data broker or someone who just stole it. It sucks 🙁
(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Karo: The most important thing companies can do is treat trust as a product requirement, not a PR problem. That means designing systems where consent, explainability, and accountability are built in from the start, treated as part of the value proposition, not added later when regulators or users push back. My first Substack article, “Fix your ethics after lunch, not after launch,” focused on exactly that idea.

Product with Attitude

Fix Your AI Ethics After Lunch, Not After Launch

a year ago · 28 likes · 9 comments · Karo (Product with Attitude)

Karen: I remember that article — it resonated with a lot of people, and “after lunch, not after launch” was a brilliant tagline!
I appreciate you making the time for this interview, Karo! Is there anything else you’d like to share with our audience?

Karo: I’d probably just add this: a lot of the most important conversations about AI aren’t happening in company-sponsored press releases. They’re happening among practitioners, builders, and writers. People like you and me, trying to make sense of it as we go. That’s why I’d encourage anyone who’s curious to spend time on Substack, where those conversations are actually unfolding.

Karen: Absolutely; we have an amazing community around the topics of AI, both in general and in . And you are one of the clear leaders here on Substack, one of the people whose articles I always look for. In fact, could you say a few words about your 2026 “Builders With Attitude” initiative?

Karo: Yes, and thank you for letting me!

It’s meant as a counter-narrative to the ‘I built an app in two hours and became a millionaire’ stories all over YouTube and X.

I wanted to show how people build real value instead of just extracting money from users. That means taking readers through the build process and the product decisions involved, including skipping features or even shelving products when they know they won’t be able to deliver on their promises to customers.

Product with Attitude

Build With Attitude ✊ - An Open Invitation To Vibe Coders Who Care About Building Things That Matter

3 months ago · 141 likes · 80 comments · Karo (Product with Attitude), Karen Spinner, Jenny Ouyang, Daria Cupareanu, Ahad Amdani, Dee McCrorey, Christina Pearl, Cristina, AviationML, Ileana, Brian Schneider, Kenny - Your Freelance Friend, Kim Doyal, Nirav Bhatt, and Priya Mathew Badger

Karen: Karo, thank you so much for joining me on this interview! I can’t wait to see what you do this year and beyond. 🙂

Karo: Thank you sooooo much for having me! Really appreciate you bringing these topics to the rest of us. We need to keep this conversation going.

Interview References and Links

Karo Zieminski on LinkedIn

Karo Zieminski on Bluesky

Karo Zieminski on Substack (Product With Attitude)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it. All works, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

📜 AISW #093: Elettra Fiumi, Swiss-based Italian-American AI cinema innovator

Karen Smiley — Thu, 15 Jan 2026 11:07:00 GMT

Introduction -

This week’s edition of AI, Software, and Wetware features a written interview with Elettra Fiumi, an 🇮🇹 Italian-🇺🇸American AI cinema innovator who is currently based in 🇨🇭Switzerland. She is an award-winning filmmaker, most recently with her AI-assisted film Mamma Robot; director, producer, editor, and AI film educator; the founder and CEO of her own film production studio, Fiumi Studios; and the author of the newsletter AI Cinema by Elettra Fiumi. We discuss:

teaching people of all ages about using AI (not always for film)
learning during her master’s in AI in business during her maternity leave, and how she keeps learning
training a Midjourney mood board or custom GPT for her business needs
how AI tools could have helped to accelerate the 10-year Radical Landscapes feature documentary on her father’s work and the 9999 project, and how NotebookLM helps her today with speaking engagements on the topic
winning an award for innovative use of AI for handling her native Florentine dialect in script writing for her latest film residency
working to establish an AI female film residency (sponsors invited!)

and much more. Check it out, and let us know what you think!

Leave a comment

This post is part of our AI6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and content. (This is a written interview; read-aloud is available in Substack. If it doesn’t fit in your email client, click HERE to read the whole post online.)

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Elettra Fiumi

Karen: I’m delighted to welcome Elettra Fiumi, a Swiss-based Italian-American, as my guest on “AI, Software, and Wetware”. Elettra, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Elettra: Thank you, Karen. It’s such a pleasure to chat with you, finally, after a long time of interacting on Substack. I was raised in Florence, with an American mother and Italian father, and spent 18 years in Italy and 18 years in the U.S. I am now based in Switzerland, indeed. I’m a filmmaker, director, producer, editor, and AI film educator. I come from an extensive background in documentary filmmaking and branded content.

I’ve worked in Europe and in the US on films, series, and branded documentaries for broadcasters and platforms like Netflix, Amazon, the BBC, the New Yorker, and others for the last 15 years. The last few years, I’ve been deeply engaged with AI as a creative and cultural tool, specifically in cinema, but also just in the creative fields and productivity.

I run a production company, Fiumi Studios. We work on traditional films, as well as AI-driven projects. I balance my time between creating crazy, experimental AI films and traditional film work. And I teach a lot, internationally, at universities, museums, film festivals, and other places. Often it’s AI cinema, but it can also be AI related to business and productivity. I teach people of all ages. I’ve taught teenagers, and I’ve taught older folks who have nothing to do with film, and people who are professional film directors.

So it’s a very interesting experience to work with AI in this kind of way, because you can also learn a lot about AI when you’re teaching AI. And it’s a great excuse, also, to continuously be in a R&D mode.

Karen: Yeah, absolutely. Learning by doing is really ideal, I think, if you can do it.

Elettra: Absolutely. I’m in a continuous state of learning. And I’ve been passionate about people and archives and memory since the first works in journalism and in filmmaking that I did. And with AI, this has become an even more pressing matter to me, exploring how we preserve cultural material, how we reinterpret it over time, and how new tools change our relationship to history, authorship, and responsibility. And I often don’t have answers to all of this, but it’s interesting to be in a continuous search of what those questions also need to be.

Karen: You’re asking some very important questions, so I think that matters more than having the answers up front. Maybe the answers need to emerge.

Elettra: Absolutely.

Karen: That all sounds really fascinating, and I’m really looking forward to learning more about what you do as we talk through this interview. Tell us a little bit about your level of experience with AI and machine learning and analytics. Sounds like you are using it professionally. I’m wondering if you’ve used it personally, or if you’ve studied the technology, formally or informally.

Elettra: My introduction to AI was during a master’s in AI for business that I did at the Italian Tech Academy based in Milan, actually during my maternity leave! And it was one of those ‘it blew my mind’ moments, which is something that you hear often when you speak to people who discover AI. They often use this mode of speech, “blew my mind”. And it’s true, that’s how I felt, and I never went back to a non-AI professional world. But it was really lacking in the audiovisual, which has always been my passion. My love for storytelling was not in that master’s program. So I continued researching and learning and whatever course I could find, or whatever YouTube video I could find. I just self-taught and also took courses from others.

And I actually continue to take courses from others, because I’m also interested in how other people are thinking about it and teaching it. And although there are some interesting online creative communities that talk about AI, such as Machine Cinema, or leading AI Cinema educators like Curious Refuge, there’s really not that much of the latest information that you can find online. It’s really behind the scenes that you can find what people are really thinking, and not just, like, trying to promote something. So I take classes and I teach classes. I am an advanced professional user, but mostly for generative AI. I don’t train foundation models from scratch. But I work hands-on with AIG tools every day, professionally, creatively, and pedagogically.

I don’t use them so much in my personal life, except for mainstream AI uses on the cell phone. I’m not so concerned with the privacy issues when it comes to the data that the AI around us gathers. But I am concerned more about sharing images of my daughter, for example, online, even though I do share. It’s sort of one of those “I’m torn” situations. I want to share what my life is like, and my life and my work are so interconnected. I don’t think of work as a foreign body. It’s just part of my everyday life, because I think about it all the time, and I’m so passionate about it, and it’s a true pillar of my identity. So it’s hard to differentiate these worlds and make these decisions as we navigate our lives, because sharing is so much a part of who I am as a person. So, yeah, it’s tricky.

But going back to your question, I design workflows that combine text, image, video, sound, research. And I often use multiple platforms together. I test a lot of different platforms to see which one works best for that specific project and need. And of course, building custom prompt architectures and research pipelines, and training models. Training a Midjourney Mood Board or a custom ChatGPT is a fairly typical day for me!

I think personalization is one of the most interesting aspects of AI tools, and it’s one of the first things I teach. It was one of the first things that blew my mind, and I think it’s also where all the more delicate issues around AI lie.

Karen: Yeah, that’s a great perspective – you have some significant experience. You also did some work in digital journalism, right? You had some studies in that?

Elettra: Yeah, I studied politics and Spanish as an undergraduate at Mount Holyoke College, an all-women’s college in western Massachusetts. And I worked in magazines for American Express Publishing for 6 years. And then the internet came along. and I felt like I needed to learn how to write for the internet, which is really funny now, thinking back on what that dream was. So that’s how I ended up at the Columbia Journalism School studying digital journalism. And on the first day, they gave me a digital camera, a DSLR, and they taught us that to be a journalist meant being able to tell stories in an audiovisual way as well.

So I was trained as a one-woman band, and got into video journalism when it was just starting. Most publications didn’t even have websites yet. From there, I started getting into documentary filmmaking, feature-length filmmaking, because the short video journalism was kind of a machine, a hamster wheel. And I wanted to dig deeper into stories and work on longer-term projects.

I started my own production company right out of Columbia, and I've had my own production company ever since. In 2011 I started my first production company in New York with another filmmaker, and six years later I started my own production company, which I still have in New York and in Lugano, Switzerland. And that has allowed me to work on any project and also transition into AI quite seamlessly..

Karen: Wow, that’s very cool. That’s a great background, and yeah, neat to know that you started out learning from your program about using a digital SLR to help you tell stories. It sounds like you’ve found your calling in telling stories with visual arts.

Elettra: Totally. I feel blessed. It’s not that easy to find something that you love. I know a lot of people that do work they don’t love. And I find it absurd, because we spend so much of our time anyway working, that we might as well do what we love, to be happy people.

Karen: That’s great. Can you share a specific story with us on how you’ve used a tool that had AI or machine learning features? I’d like to hear your thoughts about how the AI features of those tools worked for you, what didn’t work, what went well, and what didn’t go so well.

Elettra: So, there are a few examples that I love to mention when people ask me about AI tools. When I’m teaching, one of my favorite tools, which I think is largely underestimated and I’m not sure why, is NotebookLM. It’s actually a Google tool, but I think it’s maybe not promoted as much. I don’t know why. But I’ve used it in countless instances.

One of the most relevant as an example of its power is how I’ve used it for interview transcripts.

In particular, my first feature documentary called Radical Landscapes was a massive undertaking. It took me 10 years to make. It started with the death of my father, and I inherited his archives. It took me that long because I put together all of these archives and digitized everything and interviewed people in order to navigate this material, which was disorganized, and it was unclear to me what it even meant. As a result of this ambitious project, I am now an expert in this group called 9999 that was part of this movement called Radical Architecture in the 1960s. And I often continue to be invited to speak about this group and their work.

So, even though the project of the film is completed, it’s actually still a big part of my life. And I continue having to go back to this material. But of course, it’s not even that present in my memory right now, because it’s not a film I’m currently working on. So, it’s one of those things, like, you know everything, but in order to talk about it every time, you would have to go back to all the material and refresh your memory. So, I decided at one point to just feed the NotebookLM all of the transcripts. There were a myriad of transcripts. Some made the cut into the film, some didn’t. And there was a lot of extra information that is not in the film, but that is very interesting for scholars, curators, and students that are interested in that era, and in that movement, and in their work.

So, putting it into NotebookLM, which is sort of like an LLM in the sense that you can train it on specific material, and it’s smart, intelligent on a lot of different topics, but it doesn’t feed on other material that can confuse it unless you ask it to. And it can hold up to 500 sources. You can also feed it certain sources and then ask it to explore and find similar sources, like, sources that are also audiovisual, like videos, or documents, or online websites that are within that topic. And then from that, it can create a podcast of different lengths. Also, following specific instructions, it can create a mind map, which is incredible to understand and expand on very complex topics and the links between them. It can create quizzes and so much other material.

So putting all of my materials of research in NotebookLM for this very niche topic was an extraordinary experience. Hearing the podcast, created in just a few seconds, and the synthesis of all of this incredible material, was incredible. I can't believe NotebookLM was not around when I was making Radical Landscapes. It would have been … mind-blowing, going back to the mind-blowing expression related to AI.

So I think it’s useful AI when you’re talking about archives and having to dig deep into a vast amount of material that is complex. You can also, for example, select certain interviews or certain documents, as opposed to all of them, in order to get and extract a podcast or analyses of a certain set of this material. So I highly recommend it for any kind of study that you need to do.

Another similar experience of using AI in my work that has really been powerful is for a specific competition on AI script writing. I came up with this idea of a granny in Florence, my hometown. It's a comedy. And she, all of a sudden, by mistake, goes viral. And I really wanted the dialogue to reflect the Florentine dialect. Which I know, of course, having grown up there. And if I speak Florentine after a while, people can tell right away where I'm from. But if I would have to write the dialogue with a Florentine dialect in a short amount of time, because of course these competitions are also timely, it would have been very challenging. One of my favorite books of all time is this dictionary on Florentine dialect. So I uploaded the PDF of that book, and was able to write the dialogue based on that dictionary. And, in fact, I got a special mention at the competition for innovative use of AI because of this technique in that specific film.

Karen: It sounds like it was very helpful for that project as well. Is there any way in which you felt that it didn’t work well? Or, what were the weaknesses of using that, when you were making the film? Which sounds like a really fun film – like, I want to go watch these when we’re done talking here!

Elettra: I have to make it! That was just the script. So I have to make the film, but I’m glad I already have the audience! I think AI has been incredible in so many stages of the very complex filmmaking process. But where it still lacks is really at the beginning and at the end, which is the scriptwriting stage and the editing. The scriptwriting is still very banal. And it has trouble with comedy. It has trouble with emotional depth and originality. So anything that is sort of unexpected is very hard. I do a lot of experimental storytelling. And so, for me, I find that it’s great for unlocking and brainstorming ideas but not for the actual writing.

And then at the end of the process is the editing. There are some AI tools like CapCut – I'm actually in their creative partner program – that are really wonderful. Or Opus, or Descript, that can extract shorter versions of a longer film for social media cuts. They’re cool for socials but they're not at a professional quality of what a professional editor with 10 years of experience can create for more cinematic experiences. I'm not really sure what the reasons are for that, but editing is a very magical process, and it really makes or breaks the film. God, it'll be an interesting moment when there's a good AI editing tool that can create an artistic edit.

Karen: You mentioned comedy. Irony and sarcasm are sometimes very hard for LLMs. It's like when humans are learning a foreign language. My husband is always making jokes or puns in English, and his native language is Chinese. You can tell he's really fluent in American English because he can do that, you know? My Spanish and German are still pretty terrible, and my Chinese really sucks. I would never be able to do that in those languages. I think large language models have a similar struggle with not getting irony or sarcasm or comedy. It’s definitely challenging for large language models and for humans.
Since you’re based in Switzerland: There’s a new Swiss public AI tool called Apertus, the Latin for open, and they were training it on over a thousand languages and supposedly a lot of dialects. I think 40% of the content it was trained on was not in English. I’m wondering, would they pick up something as specific as a Florentine dialect, maybe? Would that be useful if it did?

Elettra: So, actually, yeah, there was a huge push to publicize this LLM in Switzerland during the Swiss AI weeks. And I was a part of it. We organized an event with USI, the Swiss-Italian University, and the Istituto Dalle Molle, a leading AI research institute in Lugano. It was a really cool moment to see a country like Switzerland take this approach. Switzerland is a very unique country because it has so many different cantons, which are sort of like regions in Italy. There's four different languages, and each canton has a very specific culture, set of laws, and language. So, it makes so much sense to create an LLM that is specific to Switzerland, so that you can really go and find out, if you're doing research on anything in Switzerland, this is a platform and a universe that is trained on this very original and unique country.

But it would not have any knowledge on a dialect like Florentine. Italy would need to do its own. And if they could get their act and wit together, then they should definitely do that, because Italy is also a very interesting country that is very new (it only unified in 1861), and has so many different dialects and cultural nuances between the regions, and even between the towns that are just, like, half an hour away from each other. Some were enemies, you know, back in the day. There’s a saying in Florence: “Meglio un morto in casa che un Pisano alla porta” (“Better a dead person in the house than a person from Pisa at the door”), and Pisa is just an hour away. So you could not go to Apertus and ask for Florentine dialect nuance, because it's trained on just the Swiss culture, language, and traditions.

Karen: Okay, yeah, that’s interesting to hear. I know they said it was over a thousand languages. They mentioned Romansh and some Romanian language variants. But I know what you mean about the regional dialects. I speak what some people call Pittsburghese from growing up in western Pennsylvania, and it’s not at all the same as people who grew up near Philadelphia, or even in the northern part of the state. I still speak it on purpose, as a matter of regional pride, I guess. I use some Pittsburghese words and phrases, even though a large language model or a grammar checker will flag them as bad English. No, I’m saying that on purpose!

Elettra: Yeah, and I think sense of humor is also so specific to a place. Like, there’s some things that are really funny to Florentines that Americans, or even Romans would just not get. Those would require people with that background training and testing the models with a lot more care and time.

Karen: So it sounds like you’ve made really good use of AI tools in your professional work, and it’s really fascinating what you’re doing there. I’m wondering if you avoid using AI-based tools for some things or for anything, and if so, if you can share an example of when you avoid it and why you choose not to use AI for those things.

Elettra: It goes back to what we were saying about AI not being up to par for the initial stages and the ending stages of the filmmaking process. Early ideation is still really important to do it in the human way. So even when I teach AI to teenagers, I always have them start with a pen and paper.

And likewise, I often come up with the best ideas when I'm driving. I love coming up with ideas when I'm driving, or traveling in a train, or on the airplane, as long as I'm not with my daughter. But you know, when I'm showering, when I'm decluttering, when I'm immersed in nature, I find that to be extraordinarily inspiring to me. And I can make neural connections in a way that is just impossible when I'm in front of the computer and when I'm being fed information and being sort of splattered in my eyes and in my brain, whether it be audiovisual or text.

And even when I'm reading a book or listening to music, I just can't. There are other creatives that instead get really inspired when they're listening to music. But if it's music with words, it kind of brings you on a specific path. I love music, but it kind of informs you in a certain way. It taints those raw idea creation spaces, at least for me.

So I try to write and maybe use an AI tool, like Wispr Flow. I love Wispr Flow. Wispr Flow transcribes immediately what you're saying, and so that saves a lot of time. If you're on the go and you don't have time to sit and write or text your idea in your notes, I just use Wispr Flow on the go, and it captures my idea. And then later, when I'm at the computer, then I can develop that further with AI tools if I want to. So that's my sacred moment that I keep human.

And then, of course, when I’m working with documentary filmmaking, and there’s sensitive interviews or sensitive information, then I avoid using AI.

Karen: Yeah, that all makes total sense, so I’m sure that your documentary subjects value your discretion and the fact that you’re respecting their privacy, and it helps for building trust with them.

Elettra: Yes, absolutely. I think it's fundamental. I mean, I haven't been working on that many purely documentary projects since entering the AI space. I did work on creating AI footage for a documentary director, and that was a really awesome project. But that was more creating the imagery based on archival images, and it was with the consent of and collaboration of the person who had done the research. And I think that it's really fundamental when you're working, not just with documentary, but with any information, that you're doing it with the collaboration of experts. Even if you have the best intentions, you might just not be doing things correctly or might not be fully cognizant of the risks or of the mistakes that you could be receiving as the output. Or just what you're putting in as an input. So having a collaboration with experts that know much more than you is fundamental.

Because as many people have noted, even if an AI LLM or other tool gives you an output, if you're not an expert on that topic, you kind of take it for granted that it all sounds right. But then, if the LLM gives me something about the 9999, and I'm a super expert on it, I can immediately tell where there's mistakes, right? It could be a glaring one, but also a nuance, which is equally bad. So it's really important to team up with the right experts for these more sensitive projects.

Karen: I've interviewed a number of students and some teachers. A lot of students say they won't use it because they don't want it to pollute their ability to do the work and to learn. And some use it selectively, as more like a tutor. A medical student told me that he and his study group won't use AI to teach them certain topics, because they don't know enough about the topics yet to ask the right questions or to know if it was leading them astray. For something that they already knew, and they just wanted to understand some connections, they could ask AI to explain it. Most of the students seem pretty savvy to me. I'm curious what you find from the students that you teach.

Elettra: Oh man, the education space is like a whole … I think we would need to write, like, an encyclopedia and a whole bunch of those pages would be filled with question marks. It's fascinating. I mean, I've taught teenagers, I've taught college-age students, and then, as I mentioned, also older folks. But talking about the younger generations: I had, in a college class, a student that was totally against it. And then he would learn and see how it could be useful, and then he would start really loving it. And then he would be really against it again. It was like a tango dance of love and hate. And other people were very curious as they would get through more and more of the lesson.

This is why it’s important to take longer lessons when you’re doing AI, and not just the short ones, because you don’t really have the time to reflect and then test and implement and explore it fully. But it allows you to really understand what the risks are, and what the benefits are, and what the power of it is. And once you understand that power, then you can make those informed decisions of, “Yes, I can use it to study in this way.”

For example, NotebookLM is fantastic to study, because you’re inputting, let’s say, the textbook in there, and then it can help you study that material in a different way. Not everybody learns by reading. Maybe they need to learn by listening, or interacting in a quiz form. If you don’t know what the different tools are, what the risks are, then it’s easier to just be afraid and say, “No, I’m not gonna use it at all.” And I think that that’s risky also for the younger generations to completely not learn AI tools and how AI works, because they will be completely excluded from the workforce.

At this point, I think we're all cognizant of the fact that AI is here, and we just need to understand how to navigate it. And we need to do that together as a community with moments of exchange and of collaboration to be able to navigate it properly. But not teaching AI to students at all, or accepting that they decide not to use it at all, I don't think that those are wise moves. But when you come to the discussion of how do you teach AI in the very younger generations, like middle school, high school, I have a lot of questions about that. I don't have an answer of how that should be done.

I also taught some architects, for example. And there, you're talking about, okay, so what do you teach the students? You don't teach them how to draw with pen and paper anymore, because you can do it with AI tools? Like, no, they need to learn those basics. Going back to what we were saying, the more you know about a topic, the more prepared you are to understand and work with what the AI tools are giving you and identify the errors.

But how do you divide that time? How much do you teach the traditional material, and when and how to introduce AI? That’s a very tricky discussion. I would love to be doing more projects around this topic, because I think it’s really of paramount importance for the future of our planet.

Karen: Yeah, absolutely, and that whole education topic is very active on Substack, where we both hang out, especially as far as use of AI in education. So, yeah, we’d have to save that for another interview, maybe!

Elettra: Yeah, count me in!

Karen: All right! So one thing I wanted to touch on with you is: we talked a little bit about feeding information into these AI tools. They also have a base training set. One concern that comes up is: where did these AI tool companies get the data that they use to train the tools on? A lot of times they’ve scraped ‘publicly available’ information, which is maybe copyrighted, or used a dataset like Common Crawl, which is copyrighted material, plus it has some awful garbage in it. So there’s a lot of questions around where they get their data. And we’ve put data into different online systems where we create accounts, and then they’ve used our data, and they’re not always very transparent about that.
There’s a concept called the 3C’s that came from CIPRI, which is the Cultural Intellectual Property Rights Initiative. And they advocate that people should be entitled to Consent, Credit, and Compensation for use of their works. As someone who creates content, as someone who uses the tools, I’m wondering what your thoughts are about the 3C’s?

Elettra: The three C’s of consent, credit, and compensation are essential. They’re fundamental. But I would add transparency. Creators deserve to know what data is being used, how it’s being used, and for what purpose. Vague language is not enough. It’s also tricky and complicated to know what the right language is because the law hasn’t caught up fully to this tech. AI systems are built on human culture, and if companies want long-term trust, they need to treat that culture with respect, clarity, and accountability, both legally and ethically. Ethically, it’s hard to put words to that, and to make them stick to that. The only thing that will make them stick to that is law.

I know an incredible AI lawyer in Italy. I try to follow her and talk to her as much as possible, and any other AI lawyers that are doing this very, very hard work right now. I think the landscape will start becoming manageable once the law catches up. And the bureaucrats that often don’t even understand AI or use it are trying to create legislation that doesn’t make sense. Yeah, you want the law to catch up, but it also needs to make sense. So that there’s enough freedom to be able to use it and also to be able to be compensated for the material that you’ve created that the companies are using as training data.

So yeah, it's multi-fold, but, just like when people ask me about using AI for documentaries, I think it just boils down to transparency. Like, be clear about what you're doing and transparent with your audience - how you used it and why. And unfortunately, a lot of the companies have not been transparent and continue not being transparent.

I know a lot of creatives. They just don’t use certain platforms, and I totally respect that. For the work that I’m trying to do, I need to know the platforms, I need to use them, I need to understand them. And it’s not fair how some of them have been trained. But they’re right now in this wild, wild west that we’re living in. I think it’s just, like, the dinosaur era of what we will see unfold in the coming years. And now it’s important to understand how everything works, and make sure to collaborate as much as possible with these companies and with the law. Lawyers and judges and everybody need to be able to find the right language and understand the mental frameworks to be able to come up with the right rules that can compensate everyone fairly.

Karen: Yeah, the one thing I hear most consistently in the 90-plus interviews I’ve done so far is that people want the companies to be transparent. I also find that some people want even, say, people who write on Substack, to be transparent about where or how they did or didn’t use AI. I’ve been encouraging everybody to write down their own AI usage ‘policy’, like, “This is how I use it, this is what I use it for, this is what I don’t use it for, and here’s what I use instead.” And that’s not something that is frozen once you’ve written it down, because as we learn, our understandings evolve, and the tools evolve. So it’s something that can change. But I think just being transparent with people does help to build trust.
I think the laws are always going to lag. That’s a big concern that I tend to hear. But there’s also, I think, some hope – and maybe it’s just wishful thinking – but there’s some hope that as consumers, we can pressure these companies into behaving more ethically, or to reward the companies that are working ethically and try to shift the balance a little bit. I wonder what your thoughts are about that.

Elettra: Yeah, when it comes to the AI video generating tools, the ones that are more ethical are considered to be Adobe Firefly or Moon Valley’s Marey 1 or Google’s Veo generators. Adobe Firefly’s outputs haven’t been that great compared to some of the others, or at least for the projects that I've worked on. But they just teamed up with Runway to integrate their model. And in fact, Runway is one of those models, Runway and Veo3, that are accepted by corporate clients for advertising projects. So that's when you know, okay, they're approved, let's say, by clients, and that means that they're sort of considered ethical because they can be legally safe.

If you're just a consumer, then I think you should just use the platforms that are ethically trained but how can these platforms actually prove this?? It would be interesting to know how these are actually trained, in specifics. Like, how long did it take you? What material did you pick and why? Where did you get the material? How did you get it? Go into detail, like, tell us fully. Otherwise you just say, “Oh, we're ethical, because of this and that.” It still remains vague to me.

But if you're a professional actually using these tools, an expert in these tools, studying and teaching them, I think it's important to use them all, and to understand them all, and see also where the art is going to be created, and what's the potential of the technology. So it's kind of a balance among all of that.

Karen: Do you have a project that you’re working on now? I’m wondering if you have any tools that you consider to be both more ethical and competent for your needs. Are there any tools that you would recommend people look at?

Elettra: I think Runway and Veo3, for example, are two that I feel comfortable using. Veo3 is just extraordinarily expensive, so for an artist, it's hard to sustain that cost. I don't have a creative partnership with Google. So, I pay, of course, the premium professional membership every month. Even if I make something with Veo3 once every two months, for me, it's still worth it compared to what the traditional production costs are.

And then also, when we're talking about ethics for filmmaking, it's also like, “Okay, so don't use AI because you're not investing in production in the traditional way, and hiring all the traditional roles.” Yes, but who's going to fund my crazy experimental films? Give me the budget, and then I'll go and do the traditional filmmaking.

The discussion around ethics is multi-fold and complex. I think fundamentally, AI has democratized filmmaking. And it’s opening a myriad of possibilities of what we can create and what we can imagine, and the stories we can tell.

I have another example that I use often. I had a student that I was teaching, a professional director. She was doing a story about the coming-of-age of a Jamaican boy. And she was trying to generate imagery to tell this story. And it was really hard to do that: either the boy was of a different skin color, or too young or too old. I told her, go to the Jamaican government and convince them that they need to fund a project to create the data set so that these tools have that diversity baked in.

And you can tell those stories, because otherwise it's history repeating itself, and you continuously can't tell the diversity in the cultural content. So that's a prime example of, what data set is omitted, and how can we make sure that the data sets are full of all representation?

Karen: That’s a really neat story. Do you know if she was successful in convincing the Jamaican government to create that dataset?

Elettra: I think she’s a filmmaker working on her films. I don’t know if she wants to go and be an activist or in discussions with the government! But I will follow up with her on that and see how that’s going, if she’s moved forward on that.

Karen: Yeah, I’m really curious as to how they would respond. I mean, it’s somewhat in the country’s interest to be represented accurately in the rest of the world, so they might be open to it? I’ve heard a lot, especially from folks in the Global South, that it’s not just a matter of pulling in more data from these other countries and other cultures, but it has to be handled within that cultural context – which kind of goes back to something you were saying at the beginning about how we handle this information, how we handle the context, and the way that we create archives.

Elettra: Yeah, the answer will never lie in just one type of role. Like, it’s not gonna be a teacher who goes and has the answers for that, or a lawyer, or a filmmaker. The ideal team behind how these models are trained and work and evolve needs to have a variety of experts.

Karen: Yeah. Do you know if any of your films, anything you've made in the past, has ever been scraped and used in training an AI tool, that you're aware of? Some people want their information to be picked up by these tools, and some people don't. So I’m curious what your experience has been, and what your thoughts are.

Elettra: Not that I know of. I don’t have any knowledge of my data being scraped. But I am also a very open book kind of policy person and creator. So, for example, people are scared about sharing their pitch decks of films. They have an idea they’ve developed, but they haven’t started making the film because they’re afraid somebody’s gonna steal their idea. I’m not of that school of thought. I’m more of the “sharing is caring”, and nobody’s gonna make that film in the same way that I will. So even if they steal that idea, it’s not gonna be how I would make it.

And in terms of getting paid for the rights of using my work, yes, I would expect that if somebody were to scrape my work, I would want to be compensated for it. And I think it would be fair to be compensated. But I'm also like, I'll move on and create other work. And it'll be great if other people can use it, and I'm flattered if they would use it. I do think it's important to be compensated and have the authorship recognized financially, because artists and filmmakers are always struggling and author’s rights are fundamental for the industry. But the industry is changing radically .. .I might regret saying this in a few years, but I think that I'm just fascinated by the technology and what it can do. And to be able to move forward with the technology, it's also important to allow for this kind of gray area as a transition.

But as I was saying, it's really important that the law catches up to be able to understand how things should be structured. If they had been structured even just 6 months ago or a year ago, I think it would have been complicated to continue evolving the technology. There's a balance between the flexibility to be able to create and the structures that need to be put in place to sustain the artists behind the work that is being scraped.

Karen: Yeah, that’s very fair. So, on the other side, as far as personal information – as consumers and as members of the public, our data is probably being used in some AI-based tools or systems. I’m wondering if you know of any cases that you could share, or if you have any concerns in this area?

Elettra: My only concern is on sharing images of my daughter, because she's so young. She's just 2 years old. And so I wonder, if I share images of her, can somebody take them and do something evil with that image, or train something on that image? Or also, what is her right that I'm not respecting? So I find it very hard to make a decision about this particular question. It's very personal. And I know a lot of parents who just always cover the eyes or the face when they publish images. But then it looks weird, it looks creepy. So I don't want to do that. And so I end up being torn. Some days I post a photo of her, and then I might not post anything for 2 months, and then I post something of her again. I'm going with the flow and following my instinct on it.

Karen: Yeah, we’ve gotten so used to sharing. I’ve seen some people that will only post photos taken from the back. So it’s still a natural picture, but you don’t see the child’s face at all. But yeah, there’s so much online exploitation nowadays, especially with girls, that it’s a hard line to walk.

Elettra: Yeah, it’s very hard to make these decisions, and even if it’s a private account. I had changed my Instagram account to private for a while. But then because everybody says, “Oh, the more followers you have, the more work you’re gonna get” kind of thing. So then I put it public again. And then, my personal life and my work life are so interconnected. I have a personal account, and I have my professional accounts. But in the end, people often hire me because they want to work with me, and they like me and how I live life and how I think about life. So I actually have many more followers on my personal account than on my professional accounts.

Karen: Oh, wow.

Elettra: So, yeah, it’s confusing. The whole social media landscape is very confusing. It’s hard to know how to handle that as a parent.

Karen: Do you know of any companies that you gave your information to that actually told you up front that they might use it for training AI or machine learning? Or do you get surprised by finding out afterwards that, you know, they buried something in the terms and conditions that says they can use it for ‘product improvement’, and they’re actually training an AI on it? How do you feel about that? Do you feel like you have a choice on whether companies use your information?

Elettra: Yeah, I think if you're really scared, you just don't use the platforms. You just don't use the tools. I've encountered situations where AI usage was very deep in terms and conditions or introduced retroactively. What complicates this is that institutions, like we were saying, legal systems, educational frameworks, cultural institutions move much more slowly than the technology itself. So everyone is expected to engage with these tools in real time, but the rules and safeguards are still catching up. And even when opt-outs exist, they often don't feel like real choices. And yeah, maybe GDPR helps in Europe, but globally, it's all fragmented. And even in Europe, I'm not really sure how much it really works, honestly. The gap between the speed and the institutional response creates a lot of uncertainty, and it's something creators are forced to navigate on their own, often without clear guidance. None of this is a very positive answer. I'm sorry to say.

Karen: It’s an honest answer – that’s what counts!

Elettra: I try. And I trust that in the coming years, that we will have more answers on how to navigate all of this.

Karen: Yeah, I’m wondering, since you’re covered by GDPR in Europe: there was a big fuss in summer 2024, where both Meta and LinkedIn laid claim to some of our content that we had in our profiles. Were you able to opt out from either of them using your data?

Elettra: I did not. I did not opt out.

Karen: But you had the option, at least? Whereas we in the U.S. did not even have the option to opt out.

Elettra: Yeah, and the option is important to have. I am ambiguous on if I want to share or not. Just personality-wise, and because of the tech innovation that I believe in, I'm like, “you know what, maybe they do need my data, too.” I have a specific profile as a woman of a certain age and background. Maybe they do need my material to do stuff with, you know? I'm not so much about privacy for myself. I am really vehemently opposed to there not being an option. There should be an option. You know, just like “woman's body, woman's choice”. You should have the choice to do what you want, and to be able to be protected in the way that you want. Always.

Karen: That’s fair! Last question: public distrust of these AI and tech companies has been growing, partly because we’re finding out about things they’re doing with our data that we didn’t know they were doing. And I think that distrust might be healthy in that sense. But I’m wondering what these companies would need to do in order to earn and to keep your trust, or if that’s even possible?

Elettra: Radical transparency and accountability are essential. But they’re not enough on their own. And what’s missing is really an early and continuous involvement of artists, educators, and affected communities in the whole development process from the beginning and throughout in a continuous way. And this involvement can’t be symbolic, or limited to advisory boards that appear just after products are already built. It should be a partnership that is there from the beginning and sustained over time, in a “let’s get our hands dirty” kind of way. And it can be technical, cultural, educational, and ethical. It needs to be treated as a holistic project. And it’s only in that case that there can be trust.

And the companies need to be willing to explain how systems work. And set the constraints and remain in dialogue as those systems evolve, because we’re talking about technology that keeps on evolving at breakneck speed. So what they said a month ago is not current. It needs to be continuously shared. It’s this kind of teamwork that I’ve mentioned already a few times, along with transparency. These two things need to be hand-in-hand.

And how this could happen would be with more moments of exchange. So, residencies, labs, conferences, festivals, official partnerships. Like, the creative partner programs are amazing, but even more in-depth and connected, and over time, and moments of talking to each other – live interaction.

Karen: Do you know of any companies that are doing this well right now? I always like to call them out if I hear of someone who’s doing a good job on this.

Elettra: What I can say is that I was in the OpenAI Sora Alpha program, and the head of artist partnerships there, Souki Mansoor – she was incredible, and I know she’s still doing awesome work there. There was just a sense of the artists really being heard, and interacting with the engineers and with the company. And that was an extraordinary experience for me. That really made me feel like we were part of what was being built. Yeah, more of that, but even deeper, would be awesome.

Karen: Great, well, Elettra, thank you so much for sharing your AI experiences with us. Is there anything else that you would like to share with our audience about what you’re doing now?

Elettra: Well, I just got back from an incredible, magical two-week filmmaking residency in Patagonia, in the Argentine part. It was under the mentorship of the director, Paolo Sorrentino of the Oscar-winning film, The Great Beauty. And I have about a month now to refine the edit. And it was my first time doing fiction and working with actors after many years of working in documentary, and then followed by this intense AI-driven filmmaking period which is still ongoing. The film that I made there is actually a continuation of another film, totally AI, that I made at a different residency in August called Mamma Robot that has been winning various awards. It’s won four awards so far in the festival circuit.

Poster for Elettra Fiumi’s film MAMMA ROBOT, which has already won four awards: AI International Film Festival: Best AI Superintelligence Film, Best Story. Ponza Film Festival: Best Experimental Film. New York International Film Awards: Best Original Story.

And I’m seeing both of these films as part of an anthology around motherhood and inter-generational dynamics, and that being a symbol of how technology and humanity interact as well. It also explores archives, which have been a passion of mine forever. And yeah, it’s really interesting to go between these AI-centered workflows through actor-led fiction, and through all of this, exploring these topics of control, power, vulnerability, and how we’re going to see things unfold with technology in the near future.

I believe that the answer to how we navigate this current crazy landscape, and how we are building today what will be tomorrow, all lies in building the relationships and the space of exchange and conversation. And because of this, I have started working on building an AI female filmmaker residency to ensure that there’s these relationships built over time, so with a mentorship program and educational opportunities for younger women as well. And so I’m hoping that I will find the sponsors to make that happen, and if not, it’ll be a residency one has to pay to attend. But I really hope to find the sponsors to create that space of exchange.

Karen: That’s great! And congratulations on the awards that your Mamma Robot film has been winning! Do you foresee it being released in the future, where we could watch it? Or what are your plans for that?

Elettra: Who knows? Yeah, I’ve been asked to post it online, but I’m holding it back until it finishes more of the film festival circuit. So hopefully it’ll be soon. But otherwise, watch out for other films coming out and being shared.

And by the way, I’m also doing an exhibition in Milan in the spring on the 9999 group that will also have some AI stuff in it.

And I’m working on a short documentary. It’s a branded documentary for a client on the 19th century Spanish painter Sorolla. I’ll animate some of his paintings with AI and integrate them with just normal live action footage, because they’re coming out with a huge catalogue raisonné of the artist’s work.

Karen: Oh, that sounds very cool. So, if someone is interested in this residency that you’re starting up, hopefully as a sponsor, but maybe also as a participant, what would be the best way for them to get in touch with you? Would it be through your Substack newsletter, or through some other way?

Elettra: Yeah, definitely through my Substack, AI Cinema, or through my email, elettra.fiumi@ gmail.com, or through Instagram, instagram.com/fiumistudios.ai. Or through you? It's easy to find me online, just google my name.

Karen: Yes, absolutely. They can comment on your interview when we publish it! You have a distinctive, very pretty name, so hopefully we can help people to find you.
Is there anything else you wanted to add? Anything I didn’t ask you that you were hoping I would ask you?

Elettra: No, Karen, you’re amazing. Keep up this incredible work that you’re doing. I love reading your Substack, and you’re one of my favorite Substackers on the platform, so keep it up.

Karen: Oh, that’s awesome to hear. Well, thank you so much, Elettra!

Elettra Fiumi’s avatar, generated with Midjourney

Interview References and Links

Fiumi Studios (fiumistudios.ai)

FiumiStudios.ai on Instagram

Elettra Fiumi on LinkedIn

Elettra Fiumi on Substack (AI Cinema With Elettra Fiumi)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

From Moon Valley https://www.moonvalley.com/marey:

“Marey is the world’s first commercially safe video model. Trained only on licensed, high-resolution footage. No scraped content. No user submissions. No legal gray zones. A foundation filmmakers can trust to protect their work and meet the highest standards of authorship.”

📜 AISW #087: Jean Gan, Singapore-based Senior Legal Manager

Karen Smiley — Thu, 04 Dec 2025 11:06:35 GMT

Introduction - Jean Gan

This article features a written interview with Jean Gan, a 🇸🇬 Singapore-based lawyer. Jean Gan is a senior in-house legal counsel with 15 years’ experience across APAC, specialising in contracts, compliance, and cross-border transactions. She is completing a Global MBA and pursuing a PhD in Law focused on AI and dispute resolution.

Beyond her corporate role, Jean is the Founder of Global Legal AI, a platform that provides research, tools, and insights to help professionals navigate AI laws and practise responsible AI governance. The Global AI & Law (GAIL) Network is the community arm of Global Legal AI that hosts panels, expert sessions and discussions to connect professionals and drive shared learning.

On LinkedIn, Jean is the founder and leader of AIgnite Women, an international initiative driving collaboration and inclusion in the AI governance space. She also runs How to Legal AI and Beyond the Clauses, where she shares practical insights on legal strategy, AI, and the future of work.

In this interview, we discuss:

using AI tools to help craft clear, concise materials as an in-house counsel
how the prevalence of AI overviews is impacting law firms and other companies who rely on internet traffic click-throughs
why she supports the 3Cs (consent, credit, and compensation)
her ideas on what AI companies can do to build transparency and trust

and more. Check it out, and let us know what you think!

Leave a comment

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Jean Gan

(1) I’m delighted to welcome Jean Gan from Singapore as my guest on “AI, Software, and Wetware”. Jean, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Jean: Hi Karen, thank you so much for having me. I’m really excited and honoured to be part of this conversation.

I wear a few hats. I’ve been a senior in-house legal counsel for about 15 years, and I’m currently completing my MBA by the end of this year. I’m also pursuing a PhD in Law at Leicester Law School, focusing on AI governance and dispute resolution.

Beyond my legal career, I founded Global Legal AI and AIgnite Women. Global Legal AI focuses on responsible AI, compliance, and governance, where I speak, host, and moderate webinars - from panel discussions and solo talks to “Ask Me Anything” sessions. AIgnite Women centres on advancing women’s leadership and equality in tech governance.

At the heart of what I do is helping organisations navigate the intersection of law, ethics, and AI in a practical way. I also started How to Legal AI, a platform designed to help legal professionals use generative AI in their work smartly, efficiently, and responsibly.

Those sound like excellent initiatives for helping people use AI constructively and ethically in law work!
(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

Jean: My experience with AI is both professional and academic.

At a professional level, I use AI in my work as an in-house legal counsel to improve efficiency in contract review, compliance monitoring, and risk analysis.

Academically, my PhD focuses on AI regulation, arbitration, corporate legal frameworks, and access to justice.

Of course, I’m always mindful of the ethical issues and potential errors that come with using these tools. On a personal level, I use AI tools daily, from drafting assistants to analytics dashboards and content creation platforms, while maintaining a cautious approach toward their risks and limitations.

(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
Reference for this question: “But I don’t use AI”: 8 Sets of Examples of Everyday AI, Everywhere

Jean: Yes, AI overviews on Google. They’re easier to access and appear right at the top of your search results. However, there’s growing discussion about how they might be affecting the content and search industry. When AI overviews summarise information instantly, users are less likely to click through to the original websites. This means publishers, law firms, and content creators could see less web traffic and fewer opportunities for visibility, even when their work informs the AI’s summary. It’s a shift that challenges how information is shared, credited, and monetised online.

Still, people value convenience and speed, especially in a fast-paced world.

For in-house counsel, using AI tools can be particularly useful because business leaders don’t want long legal memos. They want sharp, clear, and commercially focused advice. Visual tools like tables or diagrams make complex issues easier to grasp and communicate.

(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Jean: I avoid using AI for sensitive drafts or confidential contracts, especially when I’m unsure how the platform retains or reuses data. If the privacy policy isn’t transparent, I always use secure internal tools instead.

It’s also important not to outsource thinking. There’s a growing concern that over-reliance on AI could weaken our critical thinking skills. The substance should still come from you, while AI handles the labour-intensive parts such as collating data, summarising information, or organising it into tables or visuals.

(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Jean: I strongly support the 3 Cs Rule: Consent, Credit, and Compensation.

If an AI model benefits from someone’s creative or intellectual contribution, ethical governance requires that the individual’s agency, attribution, and economic value be recognised.

In practice, this means organisations should obtain clear consent before using data or creative works, provide proper acknowledgment where human input shapes outcomes, and ensure fair reward structures for contributors. These principles form the foundation of transparent and trustworthy AI ecosystems.

Without such safeguards, AI risks deepening inequality by concentrating benefits among technology owners while diminishing the value of the human expertise and creativity that make innovation possible.

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Jean: Transparency is one of the weakest points in the current AI ecosystem. Most users don’t know where training data comes from, and disclosures are often vague or overly technical.

As someone researching AI governance, I believe explainability should go beyond algorithms and extend to data provenance. Users deserve to know whose work trained the model and whether it was done lawfully and ethically.

(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Jean: Before GenAI became popular through OpenAI and ChatGPT, AI systems had already been around. Most of us have contributed to AI datasets, often without realising it. Social media posts, biometric scans, and behavioural analytics are routinely used for training. Even at airports or online verification portals, facial recognition systems collect and retain data by default. The problem is that consent is usually implied rather than informed, which weakens individual control.

Just take a look at your phone. If you’ve noticed that advertisements or sponsored posts start appearing with content similar to something you recently discussed with a friend, it’s quite obvious what’s happening.

(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

Jean: Yes, I’ve seen my data used for AI training without clear notice.

Reddit sold user content to OpenAI and Google in 2024 without direct user consent or real opt-outs. Deleting posts didn’t undo it.

LinkedIn later updated its terms to use profile and post data for Microsoft’s AI tools, with opt-outs buried in settings and unclear for many users.

Google also broadened its 2024 privacy policy to include Gmail and Docs data for AI use, offering only partial controls that limit functionality if turned off.

Across all three, consent felt more like a checkbox than a real choice.

(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Jean: While I have not been personally affected, I have seen professionals experience phishing, impersonation, and data leakage due to weak data governance. Once information is copied or sold, it cannot be retrieved. Data risk is cumulative and irreversible. Organisations must treat data as a lifecycle asset, accountable from creation to deletion.

(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Jean: Trust depends on transparency, accountability, and inclusion. AI companies must:

Disclose data sources and governance practices clearly.
Embed ethical oversight and independent auditing into model development.
Include multidisciplinary voices such as legal, ethics, social science, and user communities in decision-making.
Public trust is not a communications exercise; it is the outcome of sound governance.

(11) Anything else you’d like to share with our audience?

Jean: AI is redefining how law, business, and leadership operate. In my view, there are currently three camps: #YesAI, #NoAI, and #IDontKnowAI, made up of those who prefer to wait and see what others are doing and how this develops. Nobody is being forced to use AI, although some organisations are already making it part of their workflow.

Over time, AI skills will, in my opinion, become one of the key requirements employers look for in many industries. But, and it’s an important but, if you choose to use AI, it should be used smartly, efficiently, and responsibly.

My focus is to ensure that this transformation remains accountable, equitable, and human-centred. I currently have a full plate: I am finalising several playbooks to help legal professionals navigate the AI landscape, publishing insights on LinkedIn, writing academic articles and guest pieces, and preparing my first book on AI governance and law. Alongside this, I continue working with regional professionals, drafting white papers on corporate AI governance, and building the Global Legal AI and AIgnite Women networks.

We’ll look forward to your book on AI and governance, Jean. Thank you so much for sharing your AI experiences with us!

Interview References and Links

Jean Gan on LinkedIn

Global Legal AI and the Global AI & Law (GAIL) Network

AIgnite Women

How to Legal AI

Beyond the Clauses

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #084: Emily, USA-based computational memory researcher

Karen Smiley — Thu, 13 Nov 2025 11:06:28 GMT

Introduction -

This article features a written interview with Emily, a 🇺🇸 USA-based computational memory researcher, author of , and creator of EXD-Net (Emotional Experience Destabilization Network). We discuss:

the “heartbreak model” that she uses to “make the math more human” and help people understand EXD-Net
why she avoids AI systems based on biometric data
how confidentiality of patient data has been protected at clinical trial companies where she’s worked

and more. Check it out, and let us know what you think!

Leave a comment

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

(1) I’m delighted to welcome Emily from the USA as my guest for “AI, Software, and Wetware”. Emily, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Thank you so much for having me, Karen! I’m so excited to be here. My name is Emily and I’m a researcher. Right now, I’m working on an original theoretical and computational framework called EXD-Net which stands for Emotional Experience Destabilization Network. It uses attractor dynamics inspired by Hopfield networks to theorize how destabilizing experiences perturb memory states leading to drift, topological reorganization, and agentic-centered identity-level change.

I know that sounds a bit dense, but I try to make the work accessible and relatable by writing about it on my Substack from a humanistic angle. My goal is to help people feel the concept, not just study it. I like to think of it as “humanized computation”.

One of my most popular pieces introduced EXD-Net v1 through the lens of heartbreak which my friends jokingly call “the heartbreak model”. It’s silly but it does make it more digestible. These aren’t just abstract algorithms. EXD-Net theorizes how people remember, change, and grow. I think there’s something powerful about making complex systems feel intimate and emotionally resonant. Most people don’t care about theory unless they can resonate with it. And my hope is to make that feeling the entry point. The more we can connect science to lived experiences the more accessible and meaningful it becomes. That’s exactly what opens the conversation.

Thanks for that introduction, Emily. Can you explain your heartbreak model briefly? (And we’ll include a link to your article in the interview post.)

So EXD-Net is a hybrid theoretical and computational framework for understanding memory, specifically how memory changes in response to chaos. What this means is that it looks at how certain emotionally intensive or disruptive experiences, or what I call destabilizers, lead to adaptive learning and even identity transformation. So for example, in a breakup, if you’ve been in a relationship for quite a while, your identity often associates stability with that relationship, and it becomes a part of your emotional architecture. But when that ends, your entire foundation (which was associated with the relationship) collapses, i.e. destabilization.

And because memory is associative, everything starts to feel unstable. You wake up thinking about them unprovoked. You hear a song and you recall an old memory. Your internal landscape has been disrupted. But this destabilizing event provides the opportunity for internal reorganization. You relearn who you are, you reassociate old memories to new contexts, and you rebuild a new sense of stability for this new version of yourself.

So now heartbreak isn’t just an emotional event, it now becomes a computational event (mathematical even!). So that’s why it’s been dubbed as the heartbreak model. A bit silly and quirky, but I think it’s also fun, because it makes the math feel more human.

The Nth Dimension

Your Ex Wasn't Deep—Your Emotional Basins Are

Last week, I explored how dynamical systems offer a powerful framework for understanding our attachment styles. More importantly, why it’s so easy for us to fall back into familiar patterns. I introduced some vocabulary like attractors, basins, and scalloping…

a year ago · 5 likes · 1 comment · Emily

(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology, or built tools using the technology?

I got my Master’s in AI, and I’ve been studying the field for a good bit now before tools like ChatGPT became mainstream. Prior, I was working in a research lab in my free time analyzing latent spaces in autoencoders to understand how they modeled graph-structured learning in humans. Like mental maps or conceptual navigation. Essentially applying graph theory to neural architectures. Professionally, I was a clinical data scientist at a global medical device company. So, I’ve worked with AI both in academic research settings and real-world applications.

That’s a good overview of your professional experience. Have you used AI outside of work as well?

Yes definitely. I do use AI when I’m doing research or am learning new concepts. As much as I love scrolling on stack exchange forums, AI tools have been really helpful in condensing the amount of tabs I have open and synthesizing the information across them. It’s a great way to process when I’m working with concepts from multiple fields.

(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well? (Reference for this question: “But I don’t use AI”: 8 Sets of Examples of Everyday AI, Everywhere)

I use ChatGPT constantly. I find it really helpful to fine-tune and mirror my thoughts back to me. Sometimes I just need things worded differently to click. It’s been a great reflection tool in my thinking and learning process, honestly.

While I think it’s great for reflection, it doesn’t have the ability to offer new insights. It’s not omniscient. Which I guess is a good thing, but if I were to solely rely on what it does reflect back then there would be this huge disparity. So while it’s great to synthesize, your work doesn’t end there.

Additionally, it does get confused and sometimes it might reflect the wrong thing back. So if you’re not being cognizant, you might not even notice it, because the thing with LLMs is that they come off pretty confident even when it’s wrong, which is particularly dangerous.

Yes, there was that study earlier this year which said that 47% of the time when they are wrong, LLMs remain highly confident that they are right!
(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Some AI-based tools I avoid are biometric systems such as facial recognition. I understand there are cases where it’s inevitable, like at TSA, but in my personal usage I’m pretty selective with what tools I use and what information I provide. I find that in a world where all our information is out in the open, the one thing I do have control over is my choice in what and how I interact with these tools and platforms.

I haven’t been flying, but Tracy Bannon told me in her interview last year that we actually can opt out of TSA taking our photos and using facial recognition. But understandably, many people didn’t feel comfortable opting out, and that’s probably even more true now.
(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

I think it is something that is inevitable. AI systems and tools do require data to learn and improve. But at the same time, there are more “sensitive” information out there that I believe does need prior consent and attribution.

When your work is publicly available, or whatever you choose to share is publicly available, scraping is almost always certain. This is important because of course you would want your work to be searchable. However, the other side is that your work may also be used to train models without your consent.

In regard to this, I do believe companies should be explicit about what and how data is being collected from users, and there should be attribution when the work is being used. I think some attempts at this are when sources are cited which is a good step. However, I think there should be a more easily accessible option on the user’s side on whether they allow scraping of their work. Ultimately, this still comes down to your own choice of what you choose to publicly share online but having a scraper block or requiring prior consent and giving credit where credit is due is definitely important. You have a right to your work. That is your IP and companies should acknowledge and honor that.

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

I like to believe, for the most part, that providers aim to be as transparent as possible. When you use a product generally there are terms and conditions you have to agree to prior usage. These are there for a reason. Whether we choose to actually read it or gloss over it is another thing. I know for myself I don’t usually read the terms, but I do know that typically when you agree, there will be a section that tells you how your data will be collected and used. But just because we click accept without reading doesn’t absolve companies from being misleading. Just because something is technically stated doesn’t mean it’s actually transparent.

True! Since you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

When I worked as a clinical data scientist, our data came from the patients, the doctors, the products used, and anything else that may be related. HIPAA compliances and anonymization were obviously implemented, but everything from the case and anything that was connected to the patients’ health was tracked for up to a year or so to ensure product efficacy and safety outcomes. Of course patients are only admitted into the trials after meeting certain requirements and understanding that their data will be used and studied.

Were there any limitations on how patients’ data could be used, or by whom? Or was it generally pretty open? I can imagine that for research purposes, the companies would want to keep the use of the data unrestricted.

Patient privacy is actually very important, so not just anyone could access the data. Even when I worked directly on a specific clinical trial, I would have to request access to that trial’s data. And after a certain amount of time, we had people who would check back in to see if data access was still needed to do the analyses.

Different clinical trials had different people responsible for overlooking the databases, while data scientists would have to request access and determine what variables from the given data were important in regard to efficacy and safety analyses. It was quite a hassle at times to figure out who was in charge of which clinical trial and to request access, but it goes to show how important the company valued our patients’ data. I can’t speak for all companies, but that was how my company handled data privacy.

(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Oh, I’m sure there are so many, but facial recognition for one. Recommendation systems in our shopping habits or content consumption would be another. I personally find recommendation systems incredibly helpful when it comes to finding new shows to watch or new songs to listen to.

(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

I think it’s well known that the majority, if not all, companies nowadays definitely collect and use our data. Terms and conditions give us a sense of acknowledgement, but I don’t necessarily think we typically have much of a choice in declining if we want to use the product. And even if we were to choose to opt out, it may actually be more of a hassle more often than not.

As a data scientist, I would like to think that the data I provide is a minuscule part in the grand scheme of it all. Still, I would hope that I fly under the radar a bit when it comes to targeted data usage, but that doesn’t mean I’m unaware of how tracking works.

(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Fortunately for me I haven’t experienced any issues like such that I can recall, so I cannot speak much on that matter.

(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

I think the most important thing AI companies can do is be open with the work and research they’re doing, and what they’re learning. For example, I believe Anthropic comes off as an AI company that is more open and transparent than others. They talk about safety and alignment very early on and write extensively about how Claude is trained, and explicitly state what isn’t trained. But more so, I think the position and tone a company takes on really shows where their priority is. Anthropic’s tone and documentations make you feel like they’re focused on ethics and alignment which isn’t always the case in this field.

(11) Anything else you’d like to share with our audience?

Thanks so much for having me, Karen!

I really love how you humanize AI by showing the reality of how people actually use it in everyday life. That really resonates with me because that’s exactly what I’m trying to do with my own work. Giving it a more humanistic lens to make the complexity of it all feel a bit more intimate. It’s really nice to be able to speak about this not just as code or systems, but as things that actually affect us and how we live.

If anyone reading this wants to connect, exchange ideas, or chat about research, especially around memory, cognition, or computational modeling, I’d genuinely love that. Feel free to reach out to me directly. I love a good conversation with curious minds.

Thank you again!

Interview References and Links

Emily on GitHub (direct contact)

Emily on Substack (The Nth Dimension)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #082: Mia Kiraki, Romania-based founder & CMO at Yahini and content creator & strategist

Karen Smiley — Thu, 30 Oct 2025 10:06:33 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

(1) I am delighted to welcome Mia Kiraki from Romania as my guest for “AI, Software, and Wetware”. Mia, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Thank you so much for having me, Karen! I love the community you’ve created 🙂

In a nutshell, I’m the founder of an AI company called Yahini, which is a strategic AI content platform, and the proud author of the newsletter ROBOTS ATE MY HOMEWORK. I also occasionally build AI workflows and automations for SMBs [Small to Medium Businesses].

On my background… I hold three degrees: one in Film, one in Marketing, and one in Security & Diplomacy. So, my brain is this collision of a screenwriter’s obsession with narrative and a diplomat’s eye for complex systems. I wish I could say I’m a good detective and that Holmes would be proud of me, but that’s unfortunately not the case haha.

That’s a neat background, Mia - three distinct degrees! I’m looking forward to hearing how those backgrounds have influenced how you use AI and what you think about it.
(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology, or built tools using the technology?

I’m both a user and a builder. Basically, I train AIs and LLMs for a living.

Yahini started as an internal tool built out of sheer frustration with AI in content. My entire professional life now revolves around not just using AI, but bending it to a specific will. I’m obsessed with making it a tool for strategic thinking.

I’m always curious about how people learn AI. Did you ever have any formal education or training in AI, as part of your 3 degrees, or are you self-taught?

Self-taught! I actually hadn’t come across any formal AI training, so I’d be curious to know if there are any worth pursuing. I have a strong conviction that AI needs to be played with, not taught. You need to build things, break them, and learn from what happens. Plus, I graduated back in 2017, so a loooong time ago, when AI was probably some obscure NPC [non-player character] in a cyberpunk mobile game 😀

Do you use any tools with AI features outside of work purposes?

Yes, constantly! For multiple things:

Stock market analysis to identify red flags, compare trends…

For nutrition, optimizing micros and macros, analyzing ingredients, creating new recipes…

And probably my favourite, when I’m looking for new books or films, I use it to do a lot of research, asking for new recommendations, all the fun stuff!

And what I like most about these is that I’m not asking for one thing - I’m actually having back-and-forth conversations where I refine my understanding. Like with macros, I’ll go from “Here’s what I ate” to “How can I get the same nutrients with better bioavailability?” to “What’s a realistic meal plan that accounts for my workout schedule?”

As someone with a degree in film, have you ever tried any of the video generation AI tools?

Actually, no! But I’ve played around with animating old family photos :)

(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
Reference for this question: “But I don’t use AI”: 8 Sets of Examples of Everyday AI, Everywhere

Absolutely :) The story of the tools I build (including Yahini) is the answer to this question.

Back when I was running my boutique content agency, I tried to integrate the early gen AI tools into our workflow. And it was a bit of a disaster. We were drowning in fluff. It was creating more work, forcing us to spend hours rewriting and adding “humanity” to it.

Something went well too, though. It was a fantastic lesson in what NOT to do. It pushed me to say “Okay, this is a sin against good content. What’s the opposite of this?”

I started building my own systems and instead of replacing thinking, I aimed to structure it.

This whole process revealed to me that most AI tools were built for people looking for shortcuts. And shortcuts are boring!

Can you give a specific example of one of those frustrating incidents that prompted you to found Yahini? Like, what’s one of the tasks in your boutique content agency that you were trying to use genAI for? Which tool did you try to use? What did it do SO badly that you decided you needed to become a founder to solve it? 🙂 And how does Yahini solve that problem now?

One of the best examples was when clients kept asking how we could integrate AI into our workflow, which honestly basically meant “Can you please charge less and just use AI to write it?” 😀 That obviously wasn’t working because the output was terrible.

We tried to use generic LLMs (since back then there weren’t many content AI tools, maybe copy.ai and that was about it) to help with content creation. But, as every person familiar with AI knows, the final output was generic at best.

I realized what the problem was… we were using these tools to try to replace the writing, but what we actually needed help with was with the strategic planning, research, organizing, everything. And that’s what no tool was solving.

Now, Yahini thinks like a strategy first. It learns your business, combines that with live market analysis and applies expert-level content frameworks. You get a complete, prioritized content strategy with detailed briefs and AI agents that help any writer (human or AI) execute well.

(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Oh, constantly. I never, ever use AI for the initial, core idea.

The first moment of creation is messy, human. Fueled by a weird combo of what you had at lunch and a half-remembered dream :D AI can’t do that. It’s a brilliant synthesizer, collagist. But it has no lived experience.

I use it mostly for research, expansion, reframing, and challenging structures.

I’m also noticing that the vast majority of AI tools lack the in-depth prompt engineering needed to make creative work creative, instead of a bland template of something else.

When you use it for research, how often do you find that the results you get can be trusted to be accurate?

About 70% of the time if I use it raw, with no prior training. I always double check the sources and there are tons of instances where the AI simply interprets the results.

(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Required! It’s not even a question. It’s the only ethical way, and anything else is digital stripmining.

‘Digital stripmining’ - what an apt description!

My parents were art dealers; I grew up with the concepts of provenance and authorship. Who made it, where did it come from, etc. We should apply this in digital too.

At the same time, I like to be realistic. Even before AI, the internet allowed you to repurpose content from others, use stuff as inspiration, and a lot of things were going uncredited. This is, unfortunately, the reality we live in.

That’s certainly true - as humans, we all learn and draw inspiration from each other. And as you point out, flagrant plagiarism is not new with AI. Generative AI just put it on steroids and made it easy for anyone to plagiarize, perhaps without even realizing it.
(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Not quite… “transparency” is a bit of a buzzword for many. But also, who’s ever checked the full terms& conditions, let alone the privacy policy!

True, over 90% of people don’t read them, but when they are 20 pages of legalese, who can fault them for not reading?

If you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

Yahini is a strategic AI content platform that helps you build a ‘brand brain’ from your own unique expertise, which then acts as a strategic director for all the content you create.

You are in complete control of the source material. It learns from your public-facing content, like your website URL or articles you’ve published, and it uses that information and only that information to construct your brand’s unique POV (point of view: audience insights, use cases, pain points, unique selling propositions).

That sounds great, that users have full control of the source material. Do you train a custom GPT [generative pretrained transformer] for each user with RAG [retrieval augmented generation] to create the brand brain, or how does it work? (If that’s not revealing any proprietary or confidential information about the company)

Very good question! It’s not a custom GPT. We use RAG combined with AI agents to generate unique brand intelligence profiles for each project. It’s never a one-size-fits-all approach.

Yahini then checks that living knowledge base and consults it for every strategic decision it makes. So, basically, it never memorizes generic patterns, but actively understands what makes your business special.

(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

I think biometrics are a good example (e.g. TSA facial recognition system).

What’s interesting to me, as someone who’s studied security and diplomacy, is the trade-off we’re presented with. We’re offered a tiny bit of convenience in exchange for participating in the normalization of biometric surveillance.

Can you talk a little more about that tradeoff?

I guess what I wanted to say was that, if you think about it, what we’re really doing is participating in a massive social experiment (in this case, where the perceived benefit is saving 30 seconds at airport security).

From a security and diplomacy perspective, that means we’re accepting “convenient” solutions. Which is both a good and a bad thing, if you ask me.

(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact.

Yes, plenty of AI tools do these, especially those that you can use for free. I use these with caution and prefer to rely on my own trained systems / workflows, and using paid plans with whatever LLMs I choose to access.

I’ve never come across / found out about a company using my info for AI without consent (yet!)

There was quite a fuss about LinkedIn last September where they opted all of us in for generative AI training, by default, if we weren’t protected by GDPR. But being in Romania and the EU, hopefully you were protected?

Definitely protected! (but you never know… 😅)

If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

Always checking, always opting out if that’s the case and I don’t want my data to be used for training.

Did you opt out of AI training on Substack? How about other social media?

Didn’t opt out on Substack, because I want to be discovered by the AI search engines. But I try to opt out when I can and when it doesn’t benefit me in a meaningful way.

(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

N/A

(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Honesty!

Translate T&Cs into plain English and be transparent with them on your website’s footer, without burying them. Be blunt, let people make a choice.
Explicit and granular consent - if it’s not clear from the beginning that your company uses AI, you can ask separately on sign up: “Can we use your data to improve our service?”
Prove your provenance, and if your AI is trained on data, show me where it came from (data you’re using, APIs, etc.)

All great suggestions. Do you have any thoughts on how we can motivate the companies to actually do these things?

There are a number of suggestions I can think of, BUT I think the most important thing is the pressure that comes from the consumer…

We’re way more likely to buy from transparent companies. And when transparency actually becomes profitable, companies listen. I guess we need to vote with our wallets!

(11) Anything else you’d like to share with our audience?

I’d just like to leave people with this thought: we’re absolutely not passive victims in this. It’s easy to feel powerless against systems designed to be “opaque”. But the most effective thing you could do is out-think them.

My mission is to try to reduce the slop online and the only way to do that is to teach people, and by extension, their AI tools, to think more strategically. This is my purpose with the tools and workflows I build.

We can definitely build a smarter, more culturally-rich world, but it requires that we stop looking for shortcuts and start using our brains again. It’s a radical idea, maybe. If you’re interested in that fight, you can find me on my Substack, ROBOTS ATE MY HOMEWORK.

Until next time! :)

Love that you’re encouraging and supporting people in using their wetware with AI, Mia. Thank you so much for sharing your thoughts on AI with us!

Interview References and Links

Maria-Christina Muntean on LinkedIn

Mia Kiraki on Medium

Mia Kiraki on Substack (Robots Ate My Homework)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Subscribe now

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #078: Jeremiah Cioffi, JD, MA, USA-based attorney and AI entrepreneur

Jeremiah Cioffi — Thu, 02 Oct 2025 10:06:36 GMT

Introduction - Jeremiah Cioffi, JD, MA

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Jeremiah Cioffi

I’m delighted to welcome Jeremiah Cioffi from the USA as my guest today for “AI, Software, and Wetware”. Jeremiah, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Hello! My name is Jeremiah Cioffi, and I live just outside Washington, D.C, though I am originally from Vermont. I spent twelve years on active duty in the Army, first as a military intelligence officer and then as an attorney. I currently work as an incident response and data privacy attorney at Octillo Law, as a defense counsel in the Army Reserve, as an adjunct professor at Vanderbilt Law School, and as an advisor at an AI start-up called AkivaOS.

Wow! You sound busy 😊
What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I started using AI several years ago for personal research. A ChatGPT query here and there to get me on the right path. My use evolved when I attended graduate school at the Johns Hopkins School of Advanced International Studies. I created podcasts out of assigned readings using NotebookLM and listened to them while exercising and in the car. I also became fascinated by the game GeoGeussr, where players are dropped into a random location via a street-level panorama and must guess their position on a world map. I used LLMs to learn how to identify locations by uploading an image and prompting it to walk me through how it identifies locations.

Another great use case is resume building. I recently conducted a job search and used various LLMs to build my resumes. I would submit a complete job description and ask the LLM to provide me with key terms to use in my resume based on the description. I also used an LLM to create my current exercise plan, tailoring it with strict prompts and fitness goals.

Another use case is AI detection tools such as Grammarly and GPTZero. I use these when I suspect someone used AI to generate a document. I input the document and they give you a percentage breakdown of what percent is likely human generated and what percent is likely AI generated. The tools are not perfect - for instance, people say that em dashes indicate AI use, but I personally love using an em dash. But they are helpful, especially when the AI-generated percentage is high.

I love using em dashes too, and I see lots of writers comment that they doggedly refuse to give them up! I’m curious, what do you do when you conclude that a document was most likely written by someone using AI?

It’s an interesting question. In a context where I hope to learn either something from or about the author, I feel let down when I discover it was likely written by AI. It loses any sense of being genuine and makes me question the author’s motives. In an educational and professional setting, it further leads me to investigate whether the author violated any ethical principles.

The nuance here is that if the author is up front about the use of AI, either in the body of the text or a footnote, my feeling of disappointment disappears. In a context where the author is immaterial to my purpose - such as in a factual documentation of an event - I don’t feel quite as let down. But it does lead me to more thoroughly check underlying sources to ensure accuracy.

That makes sense. What else can you share about how you got involved with AI technologies?

I became so fascinated with AI during my master’s program that I decided to study it for my thesis. I had been depressed reading about how authoritarian regimes use AI to marginalize and surveil their citizens and set out to identify how opposition groups use AI to challenge these regimes. I identified three positive use cases centered on 2024 election cycles.

(1) In Venezuela, journalists were targeted for reporting news counter to the Maduro regime in the wake of the 2024 elections. They turned to AI, creating AI newscasters to report the news, which insulated individual journalists from regime repression and enabled them to provide a fuller picture of life in Venezuela.

(2) In Pakistan, jailed opposition leader Imran Khan and his social media team used ElevenLabs to create speeches from his jail cell. Khan passed notes to his social media team through his legal team. His social media team turned the notes into speeches and uploaded hours of Khan’s voice to ElevenLabs, resulting in speeches in Khan’s voice that they then broadcast to his followers, galvanizing them ahead of the 2024 elections.

(3) In Belarus, the opposition endorsed an AI candidate for parliament named Yas Gaspadar. He was a chatbot built on ChatGPT. While he was not allowed on the ballot, he was a source of otherwise-censored information for Belarusians who sought opposition viewpoints. These innovative uses of AI had real-world impact, primarily through creating a more open information space and protecting people the authoritarian regimes sought to repress.

Those are three great positive examples of using AI for support to marginalized people!

Once I became familiar with AI, I sought ways to work it into my professional life. This led me to AkivaOS. I started working with Awab Shamsi, the founder of AkivaOS, as an advisor because I was drawn to his vision for transforming how AI and automation can handle high-stakes, compliance-heavy workflows, and how it can do so transparently and ethically. I’ve had the opportunity to work closely with the incredible AkivaOS team and witness their orchestration of complex workflows in legal, compliance, and regulated industries, using a blend of ChatGPT, Gemini, Claude, retrieval-augmented pipelines, and human-in-the-loop review to deliver speed, precision, and security at scale.

Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
(Reference for this question: "But I don't use AI": 8 Sets of Examples of Everyday AI, Everywhere)

One of the coolest projects I’ve been involved with was advising Team Akiva as their engineers built out a Freedom of Information Act (FOIA) request automation workflow inside AkivaOS. This workflow is complex — hundreds of pages, cross-referenced exemptions, and strict deadlines. Traditionally, processing a single FOIA request could take weeks. The team designed a workflow where GPT-4, Gemini, and Claude acted as specialized AI agents within AkivaOS:

One agent classified documents based on exemption codes and sensitivity.
Another summarized findings and drafted plain-language responses.
A third agent cross-checked everything against statutory guidelines via retrieval-augmented knowledge bases.
And critically, human-in-the-loop checkpoints ensured attorneys and paralegals reviewed every decision before release.

The result? A process that used to take 40+ hours was reduced to under 6 hours — without sacrificing accuracy or compliance. It wasn’t perfect at first. Early on, we noticed that LLMs occasionally hallucinated exemption codes or misinterpreted edge-case statutes. The team quickly responded by adding structured prompt chaining, a fact-checking layer, and agency-specific RAG pipelines.

This experience reinforced something we should not lose sight of: AI’s real power isn’t in replacing people, it’s in augmenting human expertise and unlocking efficiency without losing control.

That sounds like a constructive use of AI. I’m curious about how their experiences will be with human reviews and trust of the results. On the one hand, the human-in-the-loop checkpoints are clearly essential. But there are reports though that over time, humans tend to trust the AI system more, and their oversight becomes less effective. So that’s something I watch out for, and I’m looking forward to someday hearing about an AI company handling it well!
If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

My basic rule for using AI-based tools is that I never allow the tool to have the final say. Let me illustrate this with a common example. When I research a new topic, I often use an LLM to identify and summarize on-point secondary sources. But I NEVER blindly accept the output as truth. I review the summary to get a sense of the source and then independently verify the source – both for its existence and for its content. This is an area where lawyers get into trouble. It seems every couple of weeks there is a news story about how a lawyer submitted a written product to a court with hallucinated case citations that end up being LLM generated. Not only is this unethical and lazy, but it’s also allowing the AI tool to have the final say. I don’t let AI have the final say.

That’s definitely wise. I’m curious, of all of the LLMs out there, do you find yourself preferring one tool over another for researching new topics? For instance, some of the LLMs now tout their ability to provide specific source links for their citations. Others don’t yet have it, or they may only offer it in higher-level plans.

My favorite tool is Perplexity Pro. It allows you to select from a variety of LLMs and offers three different modes: search, research, and labs. I find myself using the research mode most often. I’ve found that its source links for citations are largely accurate. I also like that Perplexity Pro has an incognito mode where my activity is not saved.

I also like the GPT-5 prompting guide (https://cookbook.openai.com/examples/gpt-5/gpt-5_prompting_guide?utm_source=tldrai). I’ve gotten better results across LLMs when I incorporate its prompting guidance. One of the tips is to prompt different reasoning effort levels depending on the task. Reducing reasoning effort limits tangential tool-calling and latency, which is helpful for simpler tasks.

As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

No - most companies are vague when it comes to training data sources. When providers say “trained on publicly-available data,” that could mean anything from open datasets to scraping copyrighted books, private blogs, or personal health forums, often without creators knowing. The lack of transparency can create legal and ethical risks, for instance users can generate derivative content that carries IP liabilities. It can also erode trust. If creators don’t know how their work is used, confidence in the tools collapses.

Absolutely. Have you seen any AI tool yet that does a good job on this? I like to call out companies that are trying to do the right thing.

While I have not seen enough transparency from AI companies, I am heartened by the movement to establish this transparency. From the European AI Act to nonprofit organizations such as the Mozilla Foundation, we are seeing both legal and public pressure to force more openness on training data. I expect to see this effort grow in the near future.

If you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

AkivaOS segregates sensitive data, secures client pipelines, and requires explicit consent for fine-tuning. They use tools like ChatGPT, Gemini, and Claude, and make sure they clearly articulate this use and explain to users how these tools train their models and use the information they input.

It sounds like your customers are providing the sensitive data and you’re taking steps to sequester it, maybe with multi-tenancy or a similar capability. So that’s good.
As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

During my transition from active military service to the civilian sector I elevated my LinkedIn game. This included providing robust amounts of professional information. It did not even cross my mind that this information could be used by an AI-based tool. It wasn’t until I saw an announcement on LinkedIn giving users the ability to opt out of having their content used to train AI that I learned the errors of my ways.

Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

About a year ago, LinkedIn gave users the option to opt out of having their content used to train AI. I have been on LinkedIn for more than a decade and never thought to dive into their terms and conditions to see how they were using my content. Once I found out, I opted out. Here is a link for folks interested in doing the same: LinkedIn and generative AI (GAI) FAQs

To LinkedIn’s credit, they make it simple to opt out - all it takes is a toggle from “yes” to “no.” However, I do not know for how long LinkedIn used my content to train AI, and they are not going back and removing my content from training data prior to my opting out. This episode solidified that I need to take charge of my content. I need to review terms and conditions and take the necessary steps to opt out where appropriate. I recommend everyone undertake a review of their social media accounts to determine if they are comfortable with how their content is being used, and to change that use where they can.

LinkedIn seems to be handling their opt-out notifications better this year. In September 2024 they didn’t handle it well at all. The default was opt-in unless you were in a country covered by GDPR or a similar regulation. And the opt-out they offered didn’t cover their use of our data up to that point in time, which they did without getting our consent. (For anyone interested in this, see my Sept. 19, 2025 LinkedIn post.)
Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

I think this is a combination of luck and good cyber hygiene. I use MFA, change passwords regularly, and spend time understanding privacy and security policies and choosing a course of action appropriately. That said, some of my information has been part of various breaches, but in each case I acted promptly to change passwords and activate provided credit monitoring services. These actions helped minimize my privacy concerns.

Those are solid practices that it would be good for anyone to follow.
Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

From my perspective advising at AkivaOS, it comes down to one thing: radical transparency. Right now, most AI companies are black boxes. People don’t know what data was used, how models make decisions, or where their information goes. That opacity fuels fear and skepticism.

Here’s what companies must start doing:

Be Honest About Data → Explain clearly what’s collected, where it came from, and how it’s used — no legal gymnastics.
Give Users Control → Let people opt in, opt out, and manage their data.
Build Accountability → Provide audit trails and explainable AI so users can trust not just the output, but the process.

This is the philosophy built into AkivaOS from the ground up. Because it serves legal, government, and compliance-heavy industries, Awab designed it with explicit consent, human-in-the-loop control, and full auditability baked in.

The companies that succeed in AI’s next chapter won’t just have the smartest models, they’ll have the deepest trust. And from where I sit, watching AkivaOS evolve, that’s the future AkivaOS is building toward.

You bring up a really good point on transparency about how models make decisions. Explainability has been a weak point for models based on neural networks. And consent - true informed consent - has been a big challenge for many companies (some might say all).
That’s all of my standard questions. Is there anything else you’d like to share with our audience?

I encourage everyone to experiment with AI tools. And do so responsibly. It will save you blood, sweat, tears, and, most importantly, time. A healthy foray into LLM use could be creating a workout plan or meal plan. Just make sure you don’t give it the final say - confirm its output with a health professional! The following link is a great place to start: https://www.heart.org/en/news/2025/03/27/ai-can-serve-up-ideas-for-healthy-meals-in-a-snap.

That’s a great suggestion, Jeremiah - and as you said earlier, always verify what a LLM tells you. Thank you so much for making the time for this interview!

Interview References and Links

AkivaOS website

Jeremiah Cioffi on LinkedIn

Jeremiah Cioffi on Bluesky

on Substack

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

6 'P's in AI Pods (AI6P) is a 100% reader-supported publication. (No ads, no affiliate links, no paywalls on new posts). All new posts are FREE to read and listen to. To automatically receive new AI6P posts and support our work, consider becoming a subscriber (it’s free)!

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #077: Sam Illingworth, Scotland-based university professor and writer

Karen Smiley — Thu, 25 Sep 2025 10:06:31 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

(1) I am delighted to welcome Professor Sam Illingworth from Edinburgh, Scotland as my guest today on “AI, Software, and Wetware”! Sam, thank you so much for joining me on this interview. Please tell us about yourself, who you are, and what you do.

Hi, I’m Sam Illingworth, a university professor and poet based in Edinburgh, where I work at the intersection of education, creativity, and technology. My background is in both science and the arts, and much of my work now explores how we can use tools like AI to create more thoughtful, human-centred ways of working. I also run Slow AI, a newsletter and growing community where I encourage people to use AI differently, not as a way of doing everything faster, but as a way of pausing, reflecting, and asking better questions.

(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I am not a developer or an engineer, but I am an experienced user, researcher, and critic of large language models. My Leverhulme-funded project is studying how students in UK universities experience AI in their learning, and I regularly write and publish on how these tools reshape education. Personally, I use AI almost daily, but always in ways that are grounded in reflection rather than productivity hype.

I know a lot of our audience members are very interested in how AI is shaping education. Can you talk a bit about your studies on that?

A lot of my current research looks at how students are actually experiencing these tools in their learning, rather than how institutions or companies imagine they should. With my Leverhulme-funded project we are collecting reflections from students across the UK about what it feels like to use large language models in their studies. Their accounts are often far more nuanced than the headlines suggest. Some students find real value in using AI as a way of clarifying or exploring ideas. Others feel excluded because they lack confidence, guidance, or even access. What this shows is that AI is already part of the learning environment, but unevenly so, and that creates issues of equity we cannot ignore.

In my own teaching I try to model reflective use. For example, I might ask AI to generate alternative framings of an assignment brief, not for students to copy, but so that we can discuss together what those framings reveal about clarity, bias, or hidden assumptions. It turns AI into a discussion partner rather than a shortcut.

That sounds like a very constructive approach! I’m also curious about how you’re using AI daily. How do AI tools impact your teaching? And do you use AI tools in your personal life as well?

Personally, I also use AI almost every day, but always with intention. Sometimes that means asking it to summarise a messy notebook so I can see patterns in my thinking. Sometimes it means prompting it for clichés so that I can avoid them in my poetry. I never fully outsource my voice, but I do use AI to create pauses that help me listen more closely to my own.

(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
Reference for this question: "But I don't use AI": 8 Sets of Examples of Everyday AI, Everywhere

One story that stays with me is when I rushed out an article pitch with the help of ChatGPT. I wanted speed, not reflection. The editor rejected it immediately and asked if AI had written it. That was a turning point for me.

The problem was not the tool, it was me skipping the pause.

Since then, I have shifted to using AI as a mirror: I ask it to challenge my assumptions, reframe my questions, or highlight what I might have overlooked. The results are far richer, and it has become a core part of how I teach and write.

That’s a great insight, and so interesting that your editor detected immediately that your article pitch was AI-generated.
(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Yes. I mostly avoid using AI to draft entire poems or creative work that I intend to publish under my name. My voice as a poet is hard-won, and I do not want to outsource that. What I will do is use AI as a provocateur: I ask it for bad lines, clichés, or metaphors to avoid. That helps me sharpen my own writing without losing ownership of the voice.

(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Absolutely. Consent, Credit, and Compensation should be the baseline. Too often, creators discover their work has been scraped and commodified without permission. As an educator and a writer, I am deeply uncomfortable with that. Companies profit from our data; the very least they can do is treat us as partners, not raw material.

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

No. Transparency is still the exception rather than the rule. Most providers hide behind vague terms like “publicly available data” without making clear what that means. For those of us working in education, that opacity is a real problem. If students cannot trust where the knowledge comes from, they cannot trust the outputs.

(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Almost certainly. I publish poems, essays, and academic articles online. It would be naive to assume none of them have been scraped into a dataset somewhere. The difficulty is we rarely know when or where. That uncertainty is part of what drove me to start Slow AI, to create a space where people can think carefully about what they share with these systems.

Your viewpoint is definitely resonating with people, because Slow AI has really taken off!

Thank you, that is kind of you to say. I started Slow AI in July, and the response has been both surprising and humbling. We are now just over 1300 subscribers, which shows there is a real appetite for using these tools more thoughtfully.

Have you ever checked LibGen or tried asking a LLM for something specific to your poetry or research articles to see if your own content comes back?

I haven’t tried LibGen, but I do sometimes give ChatGPT samples of my own writing to see how closely it can echo my voice. With prose it does a fairly good job, but with poetry it still misses the mark, which in many ways is reassuring.

(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML?
Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

The surprises keep coming. When Adobe changed its terms to allow customer images to train its models, it caught many artists off guard. I felt the same when I learned about social media platforms using posts to feed their engines. The opt-outs are often buried or absent, and users are left with little real choice. That lack of agency erodes trust.

Absolutely.
(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Not directly, but the broader erosion of trust has a personal cost. As a parent of young children, I am cautious about what family data I put online. The sense that even an innocent photo could end up training a model without my consent changes how I behave. It narrows what I feel free to share, which was already very little.

(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

They need to slow down. That may sound strange in a culture obsessed with moving fast, but building trust takes time.

Companies must be transparent about data sources,
honest about limitations, and serious about consent.

Trust will not come from glossy marketing campaigns; it will come from a willingness to pause, to listen, and to involve communities in shaping how these tools develop.

(11) Anything else you’d like to share with our audience?

Yes. I want to invite people to rethink how they use AI. The temptation is to chase the next tool or the fastest workflow. But resilience comes from reflection. That is what I write about in Slow AI, my weekly newsletter. Each post offers a simple prompt or practice designed to help you use AI more thoughtfully, whether in business, study, or personal life. If you want to explore AI as a mirror rather than a machine, that is the community I am building. Everyone is welcome.

Thank you so much for sharing your thoughts on AI with me and our audience today, Sam! And best of luck to you on Slow AI - I’m looking forward to continuing to watch it grow.

Interview References and Links

Sam Illingworth on LinkedIn

Sam Illingworth on Bluesky

Sam Illingworth’s personal website

on Substack : Slow AI

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #073: Dee McCrorey, USA-based transformation strategist

Karen Smiley — Thu, 28 Aug 2025 12:08:14 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

Karen: I’m delighted to welcome Dee McCrorey from the USA as my guest for “AI, Software, and Wetware”. Dee, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Dee: I’ve worn a lot of hats over the years, Karen, but the common thread is this: I’ve spent a lifetime navigating change — personally, professionally, and institutionally.

My roots in Silicon Valley run deep. My first job with a real paycheck was in semiconductor manufacturing at National Semiconductor while attending college in the early ’70s. Growing up in a military family, and later living abroad in Europe and the Middle East, taught me early how to adapt across systems and cultures.

When I returned to the Valley in the ’80s, I deepened the systems lens I’d first glimpsed earlier in my career — bringing that perspective into organizational change work. I became known for helping large tech companies transform from the inside out, especially during the messier post-launch phases when things went off the rails. Over time, those "triage" instincts became part of my reputation.

Over the decades, I’ve led major initiatives across the hardware and software landscape and later authored a book on innovation. Today, I write A Bridge to AI (AB2AI), a Substack newsletter that cuts through the hype to explore how AI and automation are reshaping work — and the workplace writ large —and what people can do to stay relevant and resilient.

Since launching in January, I’ve introduced Pink Slip Pivot (PSP), free monthly strategy sessions for those navigating layoffs or career transitions, which run through December.

Thanks to a bit of serendipity, in early August of AI Supremacy decided to build a Mighty Networks community in an effort to design a broader ecosystem sparking deeper connections, collaborations, debates, and shared discovery among members. He invited me to host The AI Inflection Point, a space inside the AI Vanguard Society, which I recently launched and announced to my subscribers. I believe congrats are also in order for your SheWritesAI Community that’s now part of this same network!

On the side, I’m developing a narrative podcast, Seeing Through Silicon, about the early “Wild West” days of Silicon Valley. Off the page, I’m a mom, a storytelling ‘Geema’, a mentor to local college students, and a grateful eco-volunteer living oceanside — not far from where it all began.

Karen: Thank you for that self-introduction, Dee – such a wide-ranging set of experiences! Yes, the SheWritesAI Community in AI Vanguard Society is working out beautifully so far, and I’m setting up an Everyday Ethical AI Community as well for my subscribers. I’m so grateful to for being a wonderful ally and sponsoring us, and I love that we’re building such rich network connections there. And I want to hear more about your podcast on the “Wild West” of 1970’s tech! 🙂

Dee: Seeing Through Silicon pulls back the curtain on the semiconductor industry’s rise in the ‘70s — and the predominantly female workforce that powered it from behind the scenes. Through a blend of memoir and creative storytelling, it traces how the cultural, social, and political forces of the time shaped the Valley’s values and practices — many of which still echo in today’s debates around AI and emerging technologies. By offering a more grounded lens on how we got here I hope that it will help us navigate what comes next.

Karen: That sounds so useful, Dee — it’s great that you are working on sharing that perspective. With your background in tech and Silicon Valley, I’m also wondering if you ever saw the series “Halt and Catch Fire” and if so, how realistic you think it was?

Dee: Yes! “Halt and Catch Fire” was realistic for what it was aiming to portray: a fictionalized insider's view of the personal computer revolution of the 1980s and the early days of the WWW in the early 1990s. Although it took awhile to find its place in TV lineups, the series would eventually find its fan base (including me).

Karen: My husband and I are both computer geeks and we enjoyed the show too 🙂What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

Dee: I’ve always been an early adopter of technology, but with AI, my starting point was more of a curious explorer than a technical builder — driven by an interest in its societal, cultural, and historical implications. Curiosity, though, has a way of opening unexpected doors.

Over time — and especially through Substack conversations with wickedly smart tech builders — I’ve found myself venturing into deeper waters. While I’m not a developer by training, I’m now exploring development partnerships using tools like Lovable, an AI-powered app builder, to prototype a concept I first imagined over a decade ago: a Collaborative Risktaking™ Platform. The goal? To reimagine how organizations navigate high-stakes decisions by tapping collective intelligence in more dynamic, inclusive ways.

Karen: You’ve had a dynamic career journey. I’m curious about how you made your decisions and whether any mentors helped to guide you.

Dee: I’ve been fortunate to have early-career opportunities working alongside some of the sharpest technical minds — people who helped shape my systems thinking muscle. That early mentoring laid the foundation for how I make sense of complexity today, especially at the intersection of innovation, transformation, and human behavior.

Interestingly, the dots I connected early on led me straight into the messier side of innovation — the part where things don’t go as planned. Over time, it became part of my toolkit. I developed a reputation as a kind of "tech cleanup crew" — someone who could come in post-launch and steady the ship. Eventually, both employers and clients hired me for those turnaround instincts, and honestly, that’s a pretty solid skill to have in your toolbox.

These days, my mentorship style leans into tough questions, connecting the dots between strategy and tactics, and helping others make clear-eyed decisions—not necessarily “right” or “wrong,” but aligned with who they are and what they want next.

Karen: Makes sense. I’d like to hear your thoughts on how AI is reshaping roles and career paths, and where you think mentorship can have the greatest impact—early on, during mid-career reinvention, or even later in helping others adapt or transfer skills?

Dee: Mentoring first-time and returning college students — especially amid today’s unpredictable terrain — has been deeply grounding. That experience even inspired two recent pieces I wrote on Substack, which may resonate with your readers:
• Strategic, Surgical, and Scrappy: A Career Mindset for Rough Terrain
• The AI Productivity Boom Feels Risky — But It Might Be Your Invitation to Experiment

A Bridge to AI

Strategic, Surgical, and Scrappy: A Career Mindset for Rough Terrain

Bridgework Essay | Words: 1,339 | Reading time: ~7 minutes…

10 months ago · 13 likes · 4 comments · Dee McCrorey

A Bridge to AI

The AI Productivity Boom Feels Risky — But It Might Be Your Invitation to Experiment

The Connecting Point Essay | Words: 1,482 | Reading time: ~7 minutes…

10 months ago · 22 likes · 21 comments · Dee McCrorey

Do I ever look back and wonder if being cast as the ‘cleanup pro’ was a career detour? Sure. Did some choices steer me away from the design table and into triage mode? Maybe. But in hindsight, that lens has served me well — especially now, at a moment when AI is reshaping everything.

Still, there’s risk in being perpetually cast in the damage-control role rather than being part of the design and direction from the start. And for women in tech, that pattern is all too familiar. I saw it during my semiconductor days — and I’m seeing echoes of it now in AI.

Karen: That’s an intriguing insight about women being sidelined into reactive roles in tech and AI. I don’t think many women will be surprised that it happened years ago, but it’s disappointing that it’s still happening today. Do you perhaps have a specific story you can share from your time in the semiconductor industry about a tech mess that women were tasked with cleaning up? And is there an example you can share of this happening in today’s AI world (without compromising confidentiality or anyone’s privacy)?

Dee: Absolutely. Back in the fabs, I saw firsthand how high-risk decisions — often made upstream by men in technical or leadership roles — created real downstream danger. I wasn’t a hazmat worker, but I was on the floor when toxic gas leaks and acid spills happened. These weren’t theoretical risks; they were physical, immediate, and often fell hardest on women — the majority of those working on the front lines.

That experience pushed me to move into training and eventually management, where I could influence decisions earlier in the process. I wanted those closest to risk to have a voice upstream. Sometimes it worked. But too often, decisions were still made without input from the people who had to live with the consequences.

Fast forward to today, and I see the same dynamics in AI. High-stakes choices — how systems are designed, which data is used, how automation is deployed — are often made without foresight, again mostly by men in top technical and executive roles. And once again, women are brought in after the fact — in ethics, HR, or PR — to clean up the fallout: regulatory, human, or reputational.

It’s especially troubling in HR, where AI tools are being used to justify so-called “efficiency gains” — including termination decisions made without human review. That’s a new form of harm, but the pattern is familiar: those excluded from decision-making are tasked with managing its consequences. [links below in References]

That’s why it’s critical to get women — and anyone who's ever been part of the cleanup crew — at the table before the big calls are made. People who see the whole system and ask the hard questions are exactly who we need shaping what comes next.

Karen: I completely agree! Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

Dee: I regularly use a range of AI and machine learning tools — ChatGPT, Claude, Perplexity, Gemini — for research, analysis, and ideation. I also use Notion’s AI features to build automated training and educational solutions, which has opened up new ways to deliver content efficiently without sacrificing personalization.

For communication workflows, I’m going to leverage Notion more, especially since it added a meeting recording feature that captures both sides of a meeting–a huge time saver, though I always double-check for nuance.

Lately, I’ve been exploring Mistral Le Chat, especially for building research and analysis agents. What excites me about Mistral is its commitment to open, portable, and privacy-preserving large language models. That philosophy of user-centered design and transparency really aligns with how I believe these tools should evolve.

Because I care deeply about responsible innovation, I approach AI tools with both curiosity and caution. I’m not in the “move fast and break things” camp. I believe we need strong governance and real accountability — not just for the developers, but for those funding and deploying these systems. That said, I’ve also seen how excessive bureaucracy can stall progress. So I advocate for balance — designing with intention, not just speed.

My own history in tech reminds me that bias doesn’t just show up in datasets, it shows up in who gets to ask the questions, define the problems, and build the tools. That’s why including diverse perspectives in AI development isn’t optional but essential.

Karen: I couldn’t agree more, Dee. If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Dee: Yes, I intentionally avoid entering any sensitive personal information into AI tools — things like my Social Security number, driver’s license, or birth date. I’m especially cautious when a tool’s data practices aren’t transparent or when it’s unclear how that information might be stored, shared, or repurposed.

Ironically, while I’m careful about what I input directly, many companies are already using our personal data behind the scenes — particularly in areas like hiring and healthcare. Algorithms now help screen resumes, rank candidates, and even assess “culture fit.” In healthcare, AI is increasingly used to guide treatment decisions and flag potential risk — often without patients fully understanding how those decisions are made or what data is being used.

One of the biggest concerns is the lack of clear boundaries. Where does automation end and human judgment begin? That ambiguity makes trust the real tipping point.

Who are we trusting — the developers who built the model? The company that licensed it? The vendor managing it day-to-day? Too often, those roles are siloed, and with that comes a lack of accountability.

We may not always be able to opt out, but that doesn’t mean we should opt out of the conversation. Whether it’s hiring, healthcare, or other high-stakes domains, we need to push for transparency, human oversight, and ethical alignment early in the tech lifecycle — because these systems are already shaping outcomes that affect all of us, whether we engage with them directly or not.

Karen: Absolutely.
A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Dee: My optimism on this topic swings wildly — sometimes within the same day 😣 In a better world, creators wouldn’t have to worry about their work being scraped or repurposed without consent. But that’s not the world we live in. Economic sovereignty — especially for creatives — is rarely granted. More often, it has to be fought for, clawed back inch by inch.

AI companies often justify their practices by saying they use only “publicly available” data. But as developer Ed Newton-Rex, who resigned from Stability AI over this issue, pointed out: “Publicly available” doesn’t mean anyone gave permission. It just means no laws were broken to get it. That distinction matters.

The deeper problem is how we’ve normalized trading privacy for convenience. Millions of people sign up for “free” tools and platforms without realizing what’s being extracted in return. Our rights aren’t lost all at once — they erode quietly, click by click, platform by platform.

And with recent court rulings leaning toward generative AI companies and expanding the definition of “fair use,” we’re entering what will likely be a long legal reckoning between creators and platforms.

The real question now is whether we’re willing to design systems that put human creativity, consent, and fairness at the center — or whether we’ll keep letting convenience and corporate interest write the rules by default.

Karen: All great observations, Dee. The recent rulings on fair use were a let-down for creators. But the judges’ rulings did leave the door pretty wide open for them to win damages for market disruption and for how the companies stored the content they scraped. So all isn’t lost yet. It’s definitely going to take a long time to settle out, though.
As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Dee: Why prioritize transparency when there are no real consequences for hiding the truth?

Karen: That sounds like a NO 🙂

Dee: I’d love to see AI companies lead with trust and transparency instead of power and profit — but we’re not there yet. And my deeper worry is that the pendulum won’t just swing away from transparency — it could collapse entirely. We may be heading toward a future with no real guardrails at all: ethically, operationally, or legally. And history tells us it often takes something catastrophic to force change.

What does catastrophic look like? It’s not just creative theft or misinformation. It’s militarized AI. Mass surveillance. Algorithmic control over speech, behavior, and thought. And it may not come from the usual suspects. The real danger might lie with fast-moving, well-funded startups that treat governance like a joke and see self-regulation as optional.

In a world without accountability, no one feels obligated to act responsibly. And in that vacuum, the old question applies: If a tree falls in the forest and no one hears it — did it really fall? That’s where we are with AI transparency right now.

Unless we create real systems of accountability — for those building, funding, and deploying these tools — we’ll keep mistaking silence for safety. And that’s a risk none of us can afford.

Karen: Good points - and I think a lot of people assume we have no power to make those demands for accountability, but I feel like we can’t (and shouldn’t) just give up.
As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Dee: At this point, we should all assume our personal data has been used by AI systems — often without our full awareness or meaningful consent. And given the lack of enforceable guardrails or regulatory teeth, it’s likely happening far more than we realize.

Take online job applications. Candidates are often required to hand over highly personal information just to be considered. Increasingly, that data is analyzed by AI-driven recruitment tools, which can create invisible barriers — perpetuating historical bias or filtering candidates based on flawed or opaque criteria. Most applicants never find out how their data was used or why they were screened out.

These risks aren’t theoretical. Data breaches are already happening. Bias is already baked into many systems. And the stakes are real — because access to jobs, hiring decisions, and even performance evaluations are being shaped by tools that lack transparency and oversight.

What we need is a multi-layered response: strong regulatory oversight, ethical design practices, transparent policies, and meaningful data protections. But we also need ongoing cross-sector dialogue to ensure that speed and convenience don’t come at the cost of human dignity, opportunity, or fairness.

As AI becomes more embedded in everyday life and business, protecting privacy can’t remain an afterthought. It has to be foundational.

Karen: Yeah, hiring and staffing decisions are a great example of your earlier point on how biases get baked into AI features. You may have heard about the Amazon experiment with AI for resume screening? They had to discontinue it because they realized it was reinforcing historical biases against women. [link]
I found one recent study while I was researching my upcoming book on AI ethics [Everyday Ethical AI: A Guide For Families & Small Businesses]. The researchers at Lehigh University were evaluating racial and other biases in mortgage underwriting decisions. And they found that all of the major LLMs were unfairly biased. But they also found that giving the LLMs explicit instructions to ignore race as a criterion almost completely eliminated the bias. [link] So these problems *aren’t* unsolvable. People do have to care enough about fairness to work on them!

Dee: I think your book is excellent — and so timely! I’m helping you to promote it on Substack and LinkedIn because I feel that it should be in the hands of everyone impacted by artificial intelligence, whether in their small business, on the job, or educating their children.

Karen: I appreciate that, Dee! I’m setting the launch price for the full ebook as low as Amazon will let me. I feel it’s SO important for everyone to understand what’s behind the chatbox prompts and the smart systems that we’re all using every day, whether we realize it or seek them out or not!
Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs? How do you feel about how your info was handled?

Dee: Transparent policies and informed consent aren’t just ethical checkboxes but foundational to building trust. That’s one reason I made the decision in 2024 to significantly reduce my engagement across social media platforms. I simply no longer trusted that these companies were being transparent in how they govern or use our data.

We’re seeing what happens when Big Tech is left to regulate itself. Take LinkedIn, for example. In 2024, it quietly updated its terms after reports by 404 Media surfaced that the platform — along with parent company Microsoft — had likely used member data to train AI models. In contrast, LinkedIn users in the EU, EEA, and Switzerland were offered opt-out settings, thanks to stronger regional data privacy laws. U.S. users were not.

Karen: Yes, that was an egregious example of how NOT to handle people’s consent to the use of their data!

Dee: The consequences go far beyond social media. As AI adoption accelerates in the workplace, there’s growing tension between employers eager to automate and a workforce that remains unsure — and often unaware — of how their data is being used. What’s at stake? Job security, workplace surveillance, and erosion of personal privacy.

Employers already have vast stores of employee data, and it’s not far-fetched to imagine that some will begin monetizing this information under the guise of it being “anonymous.” But anonymization doesn’t always mean protection.

Karen: Absolutely, and I think a lot of people who aren’t familiar with how anonymization works just assume — trust — that it’s effective at protecting their privacy. But it’s often not.

Dee: AI is now being embedded throughout the employee lifecycle from onboarding to performance tracking to exit interviews. Some companies are even using these tools to influence decisions about promotions or terminations, often hidden in compliance policies employees must accept as part of standard hiring agreements.

And transparency is being further undermined by legal tools like NDAs. We’re seeing a trend where hiring agreements, non-disclosure clauses, and compliance documents are being rewritten to quietly authorize AI-driven surveillance. Ironically, AI is also being used to manage those very same NDAs — automating enforcement without necessarily protecting the individual.

We’re not just facing a technological challenge but a cultural and governance reckoning as well. If we don’t intervene with stronger protections and shared standards now, the normalization of invisible monitoring and consent-by-default could become the new workplace baseline.

Karen: Those are all excellent observations, Dee. And I think we agree that these are not the kind of workplace environments we want our children and grandchildren to have to live with.
Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Dee: Over the years, I’ve received notifications about personal data breaches, though I can’t definitively trace the few phishing attempts I’ve experienced back to any one incident. Like many, I’ve dealt with the inconvenience of updating passwords and login credentials — more of a disruption than a direct harm.

But I don’t take that for granted. It feels like a matter of when, not if, especially with the healthcare industry becoming an increasingly attractive target for cybercriminals.

And when that happens, it won’t just impact individuals — it will ripple across entire communities. The intersection of AI, data security, and critical infrastructure needs far more attention before we cross into territory we can’t easily walk back from.

Karen: Do you feel like there’s still time, that we aren’t yet past the point of no return? We haven’t yet crossed into that territory?

Dee: I think we’re standing at the edge of two thresholds: a tipping point and an inflection point. How do we distinguish between the two?

A tipping point is when something snaps — often quietly — and suddenly the damage becomes irreversible. In the context of AI, data security, and critical infrastructure, a tipping point could look like a coordinated cyberattack that doesn’t just bring down a hospital or a utility — but an entire region’s ability to function. If a large-scale breach erodes public trust so deeply that people begin to avoid care, question diagnoses, or reject digital systems outright, we won’t just have a technical problem — we’ll face a legitimacy crisis.

And we’ve already brushed against that edge.

The Colonial Pipeline hack disrupted fuel supplies across the East Coast. Ransomware attacks have frozen hospital systems, delayed surgeries, and jeopardized access to life-saving data. The Cambridge Analytica scandal, the Equifax breach, and AI-driven election interference didn’t just exploit system vulnerabilities — they fractured public trust.

Each of these moments was a warning flare. High-stakes, high-impact — and still, not enough to trigger the kind of collective response this moment demands. The next breach may not offer that luxury.

But we’re not quite there — yet. Right now, we’re still at an inflection point — a window, however narrow, to shape the trajectory of trust between humans and AI. We still have time to harden systems. To increase transparency. To demand meaningful oversight. To ensure that trust is earned—not assumed.

But that window won’t stay open forever. Every breach, every flawed deployment, every opaque algorithm chips away at that fragile foundation.

That’s why I believe trust between humans and AI is the inflection point of our time.

The question isn’t just can we build better tools — but will people believe those tools are safe, fair, and aligned with human values? Because once trust is lost at scale, it’s incredibly hard to rebuild.

Karen: Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Dee: There’s a long-standing mantra in Silicon Valley: “We just make the tech — how people use it is another story.” That kind of detachment has fueled years of unchecked innovation, where proprietary algorithms scaled rapidly without meaningful oversight or responsibility.

But if tech companies want to build real trust, they need to do more than issue polished statements or go on yet another technology apology tour that fails to land. Trust must be earned through action, accountability, and a demonstrated willingness to operate with integrity, not just innovation.

The late Dr. Aaron Lazare, a leader in the study of apologies, outlined four core elements that define a meaningful one:

Acknowledgment of the offense
An explanation of what happened
Genuine remorse and humility
Reparation or justice that addresses the harm

Not every apology needs all four, but at least one must speak to the psychological needs of the offended party. And in this case, the “offended party” is increasingly everyone impacted by opaque data practices, biased models, and tech deployed without consent.

If we can't trust how systems are built, trained, and governed — if we can't trust the people behind them — then the technology itself, no matter how powerful, risks becoming irrelevant. Or worse, dangerous.

Karen: Agreed, Dee. I’m curious, regarding oversight and responsibility, what’s your take on AI regulation? For instance, do you think a federal approach to AI regulation here in the US can create a more cohesive and effective framework for ethical guardrails and accountability, vs. the patchwork of state laws we have now?

Dee: For a while, many hoped the EU’s data protection laws might set a global precedent, especially as the U.S. remained stuck debating whether AI oversight should be federal or state-led. In hindsight, that may have been overly optimistic on both counts.

Just in the last month:

The U.S. released its 2025 AI Action Plan, emphasizing deregulation, open-source innovation, and infrastructure scale.
The EU began enforcing the AI Act, tightening privacy, liability, and oversight.
China doubled down on centralized governance, linking AI access to national objectives.
And tech giants began offering $300M packages to lure top AI researchers across borders.

From a distance, these sound like policy debates. Up close, they’re reshaping how we work, adapt, and build new ways to earn a living.

Which tools your team is expected to adopt — and who gets a say
How your resume gets screened — or whether your role gets re-scoped
Whether retraining is funded — or becomes a personal burden
Whether AI decisions in your workplace are accountable — or opaque
Whether the rules around automation protect your path — or help you pivot

While the AI Action Plan isn’t why I’m launching The AI Inflection Point — a companion series to A Bridge to AI and the community that sits inside the AI Vanguard Society — it was the signal I needed. A reminder that we’re standing at a narrowing window. I believe we’re still at the inflection point, where trust, oversight, and human values can shape the road ahead — if we act now.

Karen: That’s all of my standard questions; thank you so much for sharing these thoughts! Is there anything else you’d like to share with our audience?

Dee: I’d love to invite your readers to explore The Relevance Project™ if they’re navigating change — or questioning their relevance in an AI-shaped world.

This four-part, interactive toolkit series — paired with an engaged peer community —is designed to help you reset your direction, regain momentum, and lead with clarity.

Each toolkit includes practical, actionable tools such as:
🧠 self-assessments and diagnostics
🗺️ 90-day strategic roadmaps
🤖 AI adaptation plans
🪞 leadership reflection guides

All crafted to help you respond to real-world shifts with insight and confidence.

The full series includes:
🔹 The Relevance Reset™ (featuring the Relevance Pulse Check™) - Available now
🔹 Strategic Jumpstart™ (Coming in September)
🔹 The AI Adaptation Blueprint™ (October release)
🔹 Leading Through the Shift™ (November release)

All four toolkits are included with a Premium subscription via my Substack, A Bridge to AI.

Prefer to purchase individually or as a bundle? Visit my StackShelf creator page for à la carte access and special offers launching this fall.

(StackShelf was designed by product developer, Karo Zieminski, who writes the Product with Attitude newsletter.)

Karen: That sounds like an amazing toolkit, Dee. Thank you so much for sharing your thoughts and experiences on AI with us!

Interview References and Links

References for Question #3:

AI-driven layoffs and performance reviews risk violating anti‑discrimination laws under Title VII, the ADEA (Age Discrimination in Employment Act), and the ADA if algorithms penalize protected groups. Employers remain fully liable–even if decisions are made by AI (wired.com, natlawreview.com).
The case Mobley v. Workday has been certified as a collective action lawsuit, alleging that Workday’s AI-based hiring platform systematically disadvantaged older, disabled, or Black applicants—possibly violating federal law (culawreview.org).
A recent survey found that 94% of U.S. senior managers using AI rely on it for decisions about promotions, raises, and layoffs—creating new liability risks as AI tools increasingly guide, rather than support, decisions (axios.com).

Dee McCrorey’s website

Dee McCrorey on LinkedIn

on Substack ()

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 One Year of AI, Software, and Wetware: What's Changed?

Karen Smiley — Thu, 31 Jul 2025 10:06:00 GMT

A Year?!

Since this AI6P interview series on “AI, Software, and Wetware” was announced in July 2024 and launched on August 1, I’ve interviewed over 70 guests from 21 countries (4 continents). In that time, we’ve all seen the world of AI continue to evolve rapidly. So I thought it might be interesting to re-interview my first guest, , about what has changed in AI over the past 12 months, from her perspective. (This is a written interview; read-aloud is available in Substack.)

Subscribe now

Hi Kathy - it’s wonderful to have you back on “AI, Software, and Wetware”. It was such a pleasure to kick off this interview series with you as my first guest one year ago. And I’m so curious to hear about how your experiences with AI have evolved over the past year.

Thank you, Karen. I have enjoyed following your journey and interviews with various AI users and technophiles. It was an honor to be the first person you interviewed and I’m happy to return. A lot has happened with AI in a year. I can’t imagine where we will be with AI in five years!

Right?! If we look back 5 years, it would have been very hard to predict where we are now!
For those in our audience who didn’t meet you last year, please introduce yourself - tell us who you are and what you do now!

I officially retired in April of this year, so I have a lot more time to pursue my passion of poetry and writing. I’m working on my first historical fiction novel and a Guerrilla Marketing book to take advantage of my 20+ years as an author and entrepreneur. I still publish the Indie Author Insights Newsletter and have added a section to talk about AI called Scribe Bytes.

Congratulations on your retirement! What would you say are the biggest positive change, and the biggest negative change, you’ve seen in the world of AI in the past year?

The biggest positive change I’ve seen is the rapid improvement of artificial intelligence. The biggest negative change I see is the resistance which I feel is futile. A few weeks ago, I was listening to an AI prompt engineer and eight-figure entrepreneur talk about AI. They said to embrace AI or get run over by it. That stuck with me. I prefer to embrace it as I have all technology. I still find it so fascinating coming from an analog world.

I’m not so sure it’s as black-and-white and dramatic as that. Whenever I hear those kinds of comments, especially from the 8-figure people, I always want to look at where the money is and how it benefits those people to advocate for those positions 🙂
How have these improvements in AI affected you and your clients?

AI has helped me create covers, flyers, memes, audio, merchandise, and more to help technophobic authors and entrepreneurs. I have also had people unsubscribe to my newsletter and emails when I talk about using AI. To me it’s just a tool. One that often enhances my creativity.

I remember from your initial interview that you were already using AI for lots of purposes in your business. You’re pretty up front about what you think about AI, although I’m not sure if you’ve posted a written AI policy for new subscribers?

I’m not sure what you mean about a written AI policy?

By “AI Policy”, I mean something like this article that I posted in April 2024, about what AI tools I use & don’t use and why. I see a lot more people posting AI policies now, so that people know what to expect, and they can see up front if their values are aligned.
So I’m curious about the unsubscribes. Do the unsubscribers actually say that it’s *because* of your writing about AI in the Scribe Bytes section?

Yes, a couple people told me they are not interested in anything to do with AI.

Interesting. I have to wonder if some people subscribe to you *because* of those AI tips, though!
As of July 2025, when we’re doing this interview, what is the single most valuable thing AI does for your business?

It’s hard to pick a single thing. I’ve used AI to come up with a business name, book covers, story ideas, poetry prompts, blog topics, marketing strategies, and more.

Those sound like useful applications, and I think it’s worth noting that none of those involve you having an AI tool write your book texts for you, which I think is what concerns a lot of people about using AI.
I recall that you had used an AI voice cloning tool last year. Do you still use it?

I haven’t used it in several months. I couldn’t justify the expense and some authors didn’t like using it. Also, it became more cumbersome and time consuming to create the audio as “improvements” were made to the program. I’m considering checking it out again as I’ve discovered other platforms that offer the service.

How about for you personally, how are you using AI now?

Personally, I use ChatGPT to help me pinpoint health concerns. For example: my blood test showed elevated eosinophils. I didn’t even know what that was. It’s white blood cells. I asked AI, “What’s the most likely cause of high eosinophils in someone with type 2 diabetes, lymphedema, and heart failure” which are all health issues I’m battling. It listed five possible causes that I would have never thought of including an allergic reaction to medication or parasites. It then asked if I would like a checklist to bring to my doctor appointment!

Wow, that sounds useful, and I hope the checklist helps you have a good discussion with your doctor!
As of July 2025, what is the one thing AI doesn’t yet do for you that you wish it did?

Well, it doesn’t always give me the answer I want. I recently asked which authors have the best A+ Content on Amazon. It gave me a couple good ones and an author that isn’t even on Amazon. So, I don’t always trust AI and have to double check its response.

I’m curious if you have tried Perplexity or Claude - with their recent updates, they’re supposed to provide better source references for their answers?

I did when it first came out, but not since their updates. I recently discovered an AI aggregator called Magai that offers the best AI program based on your prompt and its vast user experience database. It’s next on my list to check out. Maybe you could interview the owner?!

Good suggestion, thanks! I’ll look at the tool and I’ll be curious to hear about your experiences with it.
In hindsight, what’s the one piece of advice about AI that you wish someone had given you a year ago?

Maybe a list of which AI programs to best use for a specific purpose, and what AI companies to invest in! ;)

Ha! Yes 🙂 What advice would you like to give to writers or entrepreneurs who may be struggling with using AI?

Play around with AI and see what works for you. A lot of the platforms offer free versions. AI is not going away. In fact, it’s improving exponentially every day. AI is a tool and can be used as little or as much as you want. Authors and other entrepreneurs can save time and money using AI. Many are making money with AI. For those who can afford to hire out the various tasks it takes to publish, market, and sell a book, that’s great. And creatives who want to stay behind for art’s sake, go for it. But, if you want to compete with the ever-growing and competitive marketplace, I highly recommend you get curious about AI, ASAP!

Kathy, thank you so much (again) for your time and for sharing your experiences with us!

Thank YOU, Karen. I appreciate the opportunity to chat with you and your audience.

Your turn - poll open to all subscribers!

More to say on this topic? Please share via email reply, comment, Note, or in chat!

Leave a comment

About this interview series and newsletter

“AI, Software, and Wetware” showcases how real people around the world are using their wetware (brains and human intelligence) with AI-based software tools, or are being affected by AI. And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”.

Subscribe now

Article References and Links

Kathy Gerstorff on Substack

Here’s Kathy’s original interview:

Find links to all 70+ interviews here:

FAQ for prospective guests:

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Leave a comment

📜 AISW #067: Liz Sunshine, Australia-based documentary fashion photographer (AI, Software, & Wetware interview)

Karen Smiley — Thu, 10 Jul 2025 10:06:18 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Photo of Liz Sunshine and some of her documentary fashion photographs. Taken by Nicole Squelch. Provided by Liz and used with her permission.

Interview -

I’m delighted to welcome Liz Sunshine from Melbourne, Australia as my guest for “AI, Software, and Wetware”. Liz, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

In 2009, street style was a new concept of photography. Instagram didn’t exist, and I was just starting my career as a photographer. I had studied film in high school, but in the first year of university I watched the professional photography world rapidly shift to digital and a whole host of professional people who didn’t pivot quickly were left behind.

For the last 16 years, I have been a fashion documentary photographer, capturing what people are wearing on the streets - both everyday and at fashion weeks and events.

At the beginning of 2022, I started a project called “Our Relationship With Clothes”, a soft activism project that explores our connection to clothing through art and conversation. The project raises awareness for textile waste and encourages people to build a deeper, more mindful relationship with what they wear.

In 2023, I self-published a book called “Future Fashion” as part of the project. It’s believed to be the first AI fashion book published globally and depicts 100 AI-generated images of fashionable older women, created using Midjourney.

(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

Compared to others, I imagine I’m a very basic user. I started using Midjourney in April 2023, and by mid-July had compiled 6,000 images of fashionable older women. I decided to publish a book, and by November that year had launched in Melbourne and was sending books all over the world.

Midjourney was still in its infancy, but the images I was able to create felt real based on what I knew from my documentary work. It was emulating light patterns, fabric wrinkles, and aging bodies accurately. Quite a few of those first images went super viral to the tune of a million shares.

AI was the tool I needed at the time. It helped me create a series of visually older women that I think the world was craving, and I had been unable to document in Australia over my career - whenever I asked women of a certain age to take their picture, they always said “No, not me dear, I’ve passed my time.”

I wasn’t interested in being swept up in the hype of AI social media success, so when the project felt complete, I went back to my documentary photography and haven’t looked back until just recently. Exploring new ideas using my own images and pushing myself to create something that feels just as curious and worthwhile as Future Fashion did at the time.

Beyond AI image creation, I also use ChatGPT and Grammarly for proofreading - I’m a terrible speller and have a bad habit of skim reading my own work, often reading words that aren’t even there. I believe you only use AI for the skills you don’t want to keep, so I’m strict about using it for my writing the way I would an editor. I write a piece, then ask ChatGPT to proofread, pulling out the engaging parts or quotes, and letting me know which areas fall flat.

I also recently asked ChatGPT to reverse-engineer a colour grade on one of my old images, and help me organise my project ideas. I am very conscious of the environmental impact of AI, so only use it for tasks I can’t complete myself.

There are also applications outside my control where AI is implemented, like Google.

(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?
Reference for this question: "But I don't use AI": 8 Sets of Examples of Everyday AI, Everywhere

Following a conversation with my friend - the well-respected Australian artist Tom Blachford - I was encouraged to play with Midjourney. When I fell down the rabbit hole of creating Future Fashion, I had been trying to ‘imagine’ Australian style, pushing and pulling the technology to see whether it could emulate my documentary work. And while it couldn’t, what I saw in those first images was a really clever, nuanced depiction of reality - far more detailed and interesting than the fairy floss trees circulating on social media at the time.

It was the emulation of reality, combined with the subject matter - fashionable older women - that propelled my obsession and made me dive deeper.

Midjourney was released in July 2022, and I started creating the bulk of my Future Fashion images in April 2023, using version 5. By the time I had completed the series in July of that same year, it was using version 5.2. Working in these earlier models of AI was exciting. I made no effort to hide the fact the images were AI-generated and left oddities, like a woman standing upright with one leg. At the time, I was not only excited by what was being created but also the thought that this series of work was also documenting a period in AI that was ephemeral and fleeting.

What went well? As a documentary photographer I’m excited by discovery. While I might begin with a clear intention, so much of my documentary work - the subject, the timing - is out of my control. It’s a process that demands patience, but when the right image appears, the satisfaction is unparalleled.

Imagine waiting on a corner of a street for two hours, hoping to capture a balloon - and then, finally a man walks past holding a red one.

AI felt similar to this. Though the process was much faster, I might create 30 images in an hour that were all nothing, and then from the simplest tweak to a prompt or completely out of the blue - magic would happen. I think this sense of discovery made the process really enjoyable, and before I even realised it was happening, I had created this body of work that felt deeply true in spirit to the rest of my practice.

What didn’t go well? I think being an early adopter to anything is fraught with criticism. I sort of welcomed it and had many long conversations about why I hadn’t used real women and how — unless it’s a documentary image — most photographs are constructed, and essentially ‘fake’.

I wish people within the creative industry weren’t as fearful of the technology, and that I didn’t feel, in some shape or form, like I was creating work that embodied that fear.

(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

All the time. As a documentary photographer, I see AI as a tool that generally sits outside the scope of my professional work. Though I created 6,000 AI images using Midjourney in three months, I have probably taken 50,000 documentary images over the last three years. And as the lines of constructed and real images continue to blur, I feel increasingly drawn to film – the ultimate analogue medium – and in many ways, the opposite of AI image creation.

I also would never use AI to write a piece for me. Over the last three years I’ve been writing more and more. It’s a skill I was once very afraid of, but now value highly and would never want AI to take that away from me.

AI is an incredible tool, but it can often erode or simplify your own skillset. To me, understanding my boundaries – what skills I want to preserve, and where AI might support me – feels essential as we integrate it more and more into our world.

(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

This is hard for me to answer, as I have lived in a world of social media and image sharing my entire career - Instagram launched in 2011. While some photographers get very upset about the use of their images without permission, I’ve always been of the school of thought that this is simply the world we live in now, with social media and Pinterest continuously sharing work without the creator's consent or credit.

How do I feel about companies using my data? I’m probably quite neutral, but I do think transparency and individual control are really important. I think that we ultimately need regulation across many aspects of the world we’re moving into. And I predict that the disruption we feel now is nothing compared to what's coming.

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

No - but it was still very early days, I probably also went in very naive and didn’t know where to look.

(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

No - but this probably speaks more to my lack of education on the topic than to the absence of it happening.

(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?
Reference for this question: "Your Data is Worth Thousands—Here’s Who’s Selling It (And How to Stop It)" by

I feel like our data – similar to our attention – is the commodity of now. It’s not a question of opting in, it’s a question of how (or whether) we’re able to opt out.

(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

No - not that I’m aware of.

(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

For a long time, I think we’ve assumed that a spade is a spade. If a wool jumper is labelled ‘wool jumper’, we trust that it’s actually made of wool — when in reality, it’s often polyester (plastic) or a blend with a small percentage of wool.

In a world moving as fast as ours, distrust should be our default. It should be our self-protection mechanism - the pause encouraging us to be informed, dig deeper and educate ourselves before being swept up in a wave of new technology.

Like all the major tech companies, trust is earned over time - through consistent behaviour, clear values and real-world outcomes. I don’t think there’s a shortcut to building that trust. It’s going to take time.

(11) Liz, thank you so much for sharing your experiences with AI. Is there anything else you’d like to share with our audience?

We live in a world of constructed images. Teams of professional people layer their skills to create works that are in no way real. Models, make-up, clothes, stylists, lighting, retouching, etc, all ensure a calculated approach and outcome to the images we see every day. Even the iPhone images we see on our social feeds have been taken in ideal lighting, with a constructed scene or scenario, mood, and feeling to ensure a likeable approach to creation. Unless it is a documentary image - the images around us are not real. AI images are also constructed images, though constructed through pixels and prompted technology, at the hands of AI specialists; they are also unreal.

AI is already changing the game in some accessible and helpful ways for photographers, like generative fill in Photoshop, and some more polarising ways, like replacing studio teams with digital AI specialists to create fashion campaigns.

Like all art, photography is valuable because of what it says and who creates it, not because of the tool used in the creation. Art galleries nationwide now collect and show AI work on their walls, echoing this idea.1

The question should not be, “How will AI change the future of creatives and the world around us?” It should be, “How will you use AI in your process, and what will the work you create say about you?”

Future Fashion is available via our project website and is helping us raise money for a local poster campaign we are planning to execute in late 2025, to spread awareness for textile waste in Australia.

Liz, thank you so much for your time and for sharing your experiences with AI in this interview. Good luck with Future Fashion and with your campaign on reducing textile waste in Australia!

Interview References and Links

Liz Sunshine’s professional website: “Our Relationship With Clothes”

Liz Sunshine on LinkedIn

on Substack (my relationship with clothes)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Disclaimer: This content is for informational purposes only and does not and should not be considered professional advice. Information is believed to be current at the time of publication but may become outdated. Please verify details before relying on it.
All content, downloads, and services provided through 6 'P's in AI Pods (AI6P) publication are subject to the Publisher Terms available here. By using this content you agree to the Publisher Terms.

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

Reference links on ‘AI art’ in art galleries:

📜 AISW #059: Anonymous9, USA-based high school teacher (AI, Software, & Wetware interview)

Karen Smiley — Thu, 15 May 2025 10:06:16 GMT

Introduction

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Anonymous9

I’m delighted to welcome Anonymous9 from the USA as my next guest for “AI, Software, and Wetware”. Thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

Hello! I am a public school warrior/educator who has LOVED and excelled at his job for 33 of the last 34 years, now just a few months away from retirement. For most of my life, teaching would be the first descriptor on any attempt to nail down my identity, though I give myself strong scores for maintaining a healthy balance of life. I am a devoted husband, father, neighborhood tenant, friend, fan of virtually all of nature, thrilled to engage in many ways of moving through the natural world.

Thank you for that introduction. I love that you describe your multiple roles in life, and not just your professional identity.
(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

Honestly I minimize technological involvement to the best of my ability, so I suspect that puts me in a different realm than most of you. Early in my teaching career, I saw that it was wise to get my Master’s degree and I quickly found it most expedient to get a degree in Technology in the Classroom, just as the internet was becoming more easily accessible back in 1992, before cellphones, and classroom laptops were ubiquitous. I only bring this up because most of my classmates now run the technology department of their entire school district. Still friends with them, I used most of the classwork to argue why I considered much of technology to be hazardous to learning and student well-being. Time and experience have only strengthened that view.

I would say I have low-level familiarity with AI. Most of it is through the experience of student use of Chat GPT with or without teacher support. But I also have lots of indirect experience of it through conversations with my wife, other teachers, neighborhood friends/university professors. I suppose use of my wife’s car and technology in general means I have a relationship with AI whether I know it or not.

Yes, it’s really hard to avoid AI in daily life nowadays, even if you choose not to use it intentionally! Can I ask, how ‘smart’ is your wife’s car? Is it “connected” and does it have any AI-based safety features that you’re aware of?

Ha! That question makes me laugh. A good kind. I think the car is stupid, not smart, because it gets irked when I ignore painted lines to make more smooth curves or avoid obstacles. A few times it has decided to put on brakes when I needed to speed up. We have not really worked out Bluetooth, but the car definitely chooses odd times to suddenly play music etc. I DO have to admit, I'm starting to like the backup camera and the navigation tool. The movie “Leave the World Behind” has stuck with me, particularly the scene of Teslas smashing into each other, essentially convincing me I want to drive my own damned car, not turn it over to others.

My car is the simplest thing available. I love it. But I wish I had the old-style roll-down windows, not electronic buttons. I know people who have drowned in cars because they didn’t open the window in time.

I know what you mean about the windows - it’s odd that the auto manufacturers haven’t come up yet with a design that lets people get out when they need to. The only ‘solution’ I’ve seen is wider availability of the ‘break window glass’ devices to keep under your seat for emergencies.
I remember seeing that movie too. 🙂 My family is also still driving an older car without connected features - no backup camera or infotainment system. We hope it keeps running forever!
(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well? (Reference for this question: "But I don't use AI": 8 Sets of Examples of Everyday AI, Everywhere)

First, please know that I would never intentionally use AI, have only been led to do so through involuntary circumstances. Much as I appreciate efficiency, I am far more dedicated to creativity, work ethic, patience, collaboration. I very much want the world to slow down and consume less.

Understood. You are in good company; some of my other interview guests have expressed feeling the same way. (Example: Lis Sandi-Diaz) And, by the way, a recent study by Upwork showed that AI tools aren’t always helping people be more efficient; 77% of people reported feeling like AI is adding to their workload, not reducing it.

Last year I worked closely with a miserable student who was given the task of researching the role of AI in a topic of passion. This student HATES writing, was adamant that he would NOT blend the worlds of writing and his passions. In response I suggested he write about the role of AI in writing. We proceeded to ask ChatGPT to do things like: “Write a paper on the role of writing.” “Convert that paper into the language of an unmotivated high school student.” “Use an equal blend of research evidence and seemingly personal anecdote.” “Add more mistakes in grammar.” “End with a less confident conclusion.”

In the end, the student completed the paper, essentially narrating the quest, submitting a draft that probably saved him from failing the class. In some small way, I believe the process and my efforts helped him become a better writer. I also fear that he will use the tool more regularly even when a teacher is adamantly opposed to it.

Wow, that is a really interesting story and set of prompts. Was “the task of researching the role of AI in a topic of passion” a standard assignment given to a whole class, or was it a special assignment?

I should make clear that I am now an IA (instructional assistant), after 33 years of running the classroom. It is a HUGE step back in terms of stress and responsibility, but also in reward. I don't make the assignments, I help students navigate them.

In this case, the general prompt was for all students, and my job was to make the assignment tolerable/meaningful for them. I THINK it was the first time this teacher has given the assignment. He works in a highly collaborative group, but I see no sign the assignment has continued this year.

Ah, interesting; thank you for elaborating on that.

In another situation, I work closely with a colleague who teaches College Writing in conjunction with the local state university which means she needs to uphold their standards and curriculum as well as that of our high school. It is readily apparent that increasing distrust of students is sucking the wind out of her and she regularly gives tirades to the effect of: "It's gotten to the point that I am suspicious and sad anytime I read essays that are on time. 50% of the time, I see that work I wanted to give a strong grade to is breaking the rules, is mostly created by AI. It makes me sad and want to quit. And the truly sad thing is, the trend is spreading and I am increasingly paranoid about you all.”

That’s sad to hear that she has gotten so disillusioned by students using AI. I’m curious about school policies on allowing students to use AI, and whether AI-based tools are used to detect use of AI or to detect plagiarism. It sounds like your colleague’s university may not be strict about it. How is it in your high school?

Warning, my sarcasm, distaste, and reasons for retiring may come through. From my perspective, there are no meaningful norms any more, no policies that are consistently upheld, no real attempt to hold students accountable. The bottom line is that no kid should be held back from graduating on time. Every teacher I know is absolutely fed-up and struggling to try, to care. As has always been the case, teachers are too busy to really collaborate and exchange meaningful policies and so there is hardly any consistency on anything.

I don't consider this teacher disillusioned. She is aware and she cares and so she tries hard to thwart inappropriate use. But that’s like one traffic officer trying to uphold traffic rules for a whole city. We have policies in print, but they are pretty meaningless in the end, like the “no camping in parks” signs next to every encampment in town. I suspect the University has similar rules in print, similar lack of ability to enforce. My neighborhood is mostly professors in that University and the power of “likes” and popularity surveys is wildly impactful. Students don’t give professors “likes” for upholding policies regarding AI.

Got it … thank you for sharing your experiences with this.

Last one (though I have many more). Recently we happened to travel through the hometown voice-artist who is a staple source for my wife’s company. He took us to a nice dinner, thanked us profusely with the words “I don't think I could live this life if I didn't have your company as a client.” He was an absolutely wonderful man. Less than a week later, the company’s founder announced an experiment, to secretly use AI voice artists in the work to see if anybody noticed or cared. Nobody noticed or cared and the company has since stopped using the human voice artist. I consider this an abomination.

Oh, that’s awful. I’ve read about so many creative and talented people who are losing their livelihoods because of people using AI tools. One of my earlier interview guests [Anyela Vega] reported that a friend who was a voice artist had his voice literally stolen, by people who made an AI clone of it without his consent; and now they get income from voice jobs with their illegal clone of his voice, instead of him. That’s so unfair.
I would hesitate to even call the clones “AI voice artists”. They’re not artists. They are just clones.
I did some research on AI voice cloning last summer. I found only a handful of ethically-trained voice cloning tools that let people clone their own voices, but not other people’s. There are dozens of unethical AI voice cloning tools. A few companies had success in taking down some of the unauthorized voices, but there are still so many out there.

I hate to say it but that sounds right, just how these things work. Record and use a distinctive voice. As you know, we have many rules and regulations that are ignored or mocked. And then we have realms that we have yet to regulate.

(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

I have many. Essentially, if I know AI is involved, I turn away. As a newcomer to Substack, I won’t read anything illustrated by AI. If I can avoid a robot phone system, I will.

I DO use navigation apps, and occasionally check out what Spotify algorithms (and similar systems) find for me, but try to avoid it.

Last year, I mentioned the challenge and time-consuming nature of writing letters of recommendation for all the students that request one to one of my administrators. They immediately suggested I use ChatGPT to write the letters, and said that they frequently do so. I lost much respect for the administrator, and made clear that I will never follow that example. I am horrified by anyone who does this.

Ugh. That is disappointing. And there’s the risk that the more often people send in AI-generated letters of recommendation, the less value letters of recommendation will have.

Grades have lost their value for lots of reasons, as have SAT scores. Students are writing (and reading) less. We have been advised to be less personal in recommendations, sticking only to attendance and grades. Personal essays are increasingly written by AI. Hard to say what matters any more.

I mean, if the recommendation letters are supposed to only cover attendance and grades, we definitely don’t need AI for that - just some simple automation that pulls from the school records. That would be about as much value as corporate verifications of employment are, nowadays, in some places; leaders are sometimes told that dates of employment and job titles are all the company will verify. A lot of them use a third party service. And I’m confident that those employment verifications are automated - and not done with AI.
(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Absolutely, most definitely, they need to get consent, give credit and compensate more than they imagine. Because of AI it is only more challenging for musicians, artists, writers etc to make reasonable living from their work, and it is utterly grotesque that publishing work feeds and strengthens the power of the Enemy. I have heard and believe that Spotify is working on AI-created music to feed new music based on people’s listening history. This is staggeringly gross, makes me want to abandon Spotify. Perhaps I will.

Yes, Spotify has done some interesting things with ML for music recommendation algorithms that help people find music they like, which seems ok. But reports started coming out in late 2023 that they’ve also started using AI-generated music [ZDnet article]. And I’ve heard from musicians that their own music is being crowded out on streaming platforms by AI-generated songs. Maybe I should un-sync my Substack podcast from Spotify.

I know beyond the shadow of doubt that what you say is true. Can go on and on and on about it, but then raise my blood pressure too high. I was super proud of Neil Young and Joni Mitchell for steering clear of the platform, but they too have caved. I have yet to cut my ties, but believe I should.

My antidote is going to lots of live music, ALWAYS buying a recording of their music from the merch table rather than through the giants. I have posted a few Spotify recommendations on Substack and believe I lost a couple subscribers due to apparent hypocrisy. That, of course, is not the issue. The ease of sharing work without consulting the artist is of course problematic.

I’ve seen some articles saying that the flip side of so many AI-generated songs is that live music performances will rise again in value. And I can see that happening. But I recall that when I first started looking at uses of AI in music last year, one musician flagged the use of in-concert AutoTune as an abomination! And universities are looking at using AI to ‘improve’ real-time AutoTune. So even a live performance might not be AI-free … but I’m sure many music lovers would still prefer it to lists of AI-generated songs. It’s great that you are supporting live music and artists that way!

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

I try not to use the tools. Whether or not I use them, I do NOT believe or feel that the process of training AI is appropriate, transparent. Sadly, the pattern of using others' work without consent, credit, or compensation is spreading.

An example of this is the large number of posts on Substack (and other platforms I presume) where people simply post pictures or art they have found without any added value or information. It is actually rare that credit for a photo is given. This is flat out theft. The more it happens, the more it will continue to happen.

I agree. I try to make sure I give credit on every image I use, and it doesn’t take that much time. I don’t get why people don’t give credit.

If you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

The closest I can come is a neighbor who works in robot science at the local university. Much as I love him, I am frequently furious with him for failing to see the repercussions of science and technology. He had absolutely no concern about robots taking over the warehouse jobs around here, saying “They are bad jobs anyway” and continually professing that “Science will fix the problem” no matter the extent of the problem.

It sounds like he’s quite the optimist. The job impact you’re describing is one of the “top five AI ethics concerns” I identified in my March articles. It seems like many people are unaware of job impacts or other harms until someone close to them gets hurt. I do agree with him that science can help to fix many problems, but helping displaced people transition to new jobs is not up to science. It’s up to our society, and it’s not clear that our society is taking the scope of that challenge seriously yet.
(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

I intentionally stick my head in the sand and avoid AI context to not be able to provide anything useful here.

Fair enough! Staying ‘off the grid’ as much as possible has its merits, for sure.
(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact.

I have likely been irresponsible, not trying to pay closer attention. This series of questions and concerns only further validates my instinct to steer clear of technology as much as possible.

Have you felt like you had a real choice about opting out, or about declining the changed T&Cs? How do you feel about how your info was handled? (Reference for this question: "Your Data is Worth Thousands—Here’s Who’s Selling It (And How to Stop It)")

Similarly, this validates my stance against using Social Media. As indicated before, I am a fairly new user of Substack and love many elements of the platform, but am getting closer to dropping out. [Note: since this interview was written, my guest has indeed left Substack.]

Yeah, the pros and cons of sharing our data and using social media are something we all have to weigh for ourselves.
For instance, I’ve decided to trade off some privacy because I’ve committed myself to speaking out about AI ethics and inclusion. But I still want to protect my privacy as much as possible. So I turned off the AI training feature in Substack; I offer my guests the option of being anonymous; and I set up a subscriber-only chat so people can ask & discuss questions that aren’t public. Those are my tradeoffs. The other part is just being vigilant about any future changes in policies and settings options … that’s ongoing.
(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

I regularly receive texts and emails indicating that security walls have been breached, that personal information has been compromised. I believe some are real and some are phishing. They all terrify me, not individually, but en masse. Increasingly the world seems run by hackers and thieves.

Yeah - one previous guest pointed out (accurately) that this problem with misuse of our data predates the prevalence of AI. The risk with AI is that it makes some scams much easier and more effective. Like voice cloning - there are stories of scammers using an AI clone of someone’s voice to try to get their relatives to send money. I feel like the more we make people aware of these risks, the better they can protect themselves. So that’s part of why I do what I do, and write what I write, and why I do these interviews.
(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

I have no idea what they can do. I believe this whole thing is a mirror of human nature. Personally, I don't need to work less or less hard. I consider work as an honor and duty. I believe this is a horrendous example of the cyclical power of expectations. I hope that was sufficiently clear.

(11) Thank you so much for sharing your time and thoughts about AI and data. Is there anything else you’d like to share with our audience?

Good luck out there.

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that all content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊
Share

📜 AISW #058: Grace Shao, Hong Kong-based AI consultant (AI, Software, & Wetware interview)

Karen Smiley — Thu, 08 May 2025 10:06:20 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview -

I’m delighted to welcome from Hong Kong as my guest for “AI, Software, and Wetware”. Grace, thank you so much for joining me today! Please tell us about yourself, who you are, and what you do.

I started my career as a tech and business reporter. Over the years, I’ve made a few pivots, but I have come full circle now, running AI Proem and focusing on AI and tech developments. With the increasing information gap between Western audiences and China, I spend a fair amount of time trying to help decipher China’s AI and tech space in a nuanced way for my English audience, which is mostly investors and industry professionals.

I am a tech and AI analyst and writer. In addition to writing the newsletter AI Proem, I often write about AI, technology, and corporate governance for Fortune, The Diplomat, Diplomat Risk Intelligence, Economist Intelligence Unit, and FT Chinese.

Beyond my editorial work, I am the founder of Proem Communications, where I work with AI and tech companies on their international positioning or media training needs. I have advised clients such as Lenovo, Ant Financial, Kuaishou, PayPal, Alibaba, AI companies, and growth-stage consumer tech and biotech firms.

I grew up in Canada and Mainland China, but I have also lived / worked in Singapore, the US, and Hong Kong. I enjoyed traveling a lot before the kids came and was fortunate enough to visit over 45 countries before a more settled-down lifestyle now.

I’m curious about the “Proem” name you chose for your newsletter and your company. It’s not a commonly used word in English, so many people in our audience may not know it. (The dictionary definition says it means ‘preliminary comment’ or ‘prelude’.) May I ask, how did you come to choose the “Proem” name for your newsletter and your company?

I really liked the word “preface” and I wanted to use that at first, but that word was quite commonly used in businesses, and in fact, in Hong Kong there is a coffee shop called Preface that is quite popular and I think an IT consulting firm I found on Google search as well under Preface. So, I thought for a long time that I liked the simplicity of a single word, and I looked up similar words to “preface” to find the word “proem.” I was a bit worried that the word isn’t commonly known, but now I kind of like that; it’s a bit geeky and nerdy. And people ask about it, just as you have. So it’s an opportunity for me to tell the story.

It means the introduction and to me for the newsletter AI Proem - I wanted to show people that I was also on a learning journey, and they could follow along to learn the basics and dive deeper as we progress. For the consulting side, Proem Communications - I thought it flowed well too, as the work I do is mostly advising companies on how to prevent or mitigate crises and conflicts, so it’s good to be prepared and “know the basics”.

That’s a cool origin story! Thank you for sharing it.

What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I cover the development of AI mostly from a business perspective. I combine my experiences working in tech strategy and journalism and bring analyses of industry trends, business models, and the impact of tech development on society. I don’t do many product reviews, but I’m always open to learning more about new products on the market and new vertical integration of AI.

Do you use any AI-based tools personally, outside of your work?

Not as much, as I’m quite mindful of the excessive energy it uses up for each query, so if I can just use a simple search for an answer, I try to still use search engines.

Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

I use Perplexity, OpenAI Deep Research, DeepSeek, and Kimi for my research. I admit, they’ve made research so much more efficient. However, I am still quite wary of the factual accuracy of some sources the chatbots cite, so I usually double-check the original source and conduct traditional desktop research when writing my pieces.

I wasn’t familiar with Kimi - that’s interesting to learn about it. How do you decide which of those four tools to use, when, for your research?

Kimi is made by a Chinese AI company, so I use it for when I research local Chinese material or need help with translation copyediting.

I still find Perplexity the best at research and I often double-check sources and fact-check with ChatGPT. DeepSeek is just another alternative, but I don’t use it as often.

I have not been able to find a really good image-generation tool yet, so if you have any tips, please send them my way.

Unfortunately, there aren’t yet any image generation tools that I’d be comfortable recommending, based on the ethics of how they’ve been developed! But if I hear of one, I will definitely let you know.
The ‘most ethical’ at present seem to be Adobe Firefly and NVIDIA Generative AI (link). I’ve read positive comments about Canva’s ethics, but their partnerships and their acquisition of Leonardo have raised some concerns (link1, link2, link3; here’s one guide to using Canva legally).

If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

I think I try not to let it write for me. However, I do occasionally seek editorial advice. I do not depend on the AI tools for full-text drafts, but I sometimes prompt the bot to tell me how I could improve the flow of a few paragraphs. I find it quite useful. It’s almost like having a copy editor, and if you prompt the programs right, you can even have an editor checking to see if you’ve missed key arguments or challenge you. Essentially, having an editor at your disposal.

I have obviously had editors at my full-time journalist roles, and they were wonderful to 1) offer guidance on pieces I might be missing, 2) provide fact-checking, and 3) copyedit grammatical errors. The beauty of a human editor is in the human interactions, the ability to talk through nuances or misunderstandings. The challenge is that sometimes they may carry biases and influence your coverage. An AI tool, in many ways, can do the three mentioned tasks, but it doesn't challenge your judgment as much, I guess. In a way, it’s beneficial; I feel like my editorial decision is then solely mine. However, on the other hand, it’s always good to have your biases checked.

That’s a good insight - thank you.
A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get consent from (and compensate) people whose data they want to use for training?

Everything is on the public internet, so it is very hard to control, especially when we join the platforms. We’ve all agreed that they can use our data. That said, as a trained journalist, I always say to give credit or link sources if the information is not first-hand.

Fair enough :)
As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Not the technicalities of how they obtain the data.

As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Not familiar with any.

Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

I think if the companies can be more transparent with how they collect data and utilize it for training and how that may impact the outcome of information (how it could be biased), it would be immensely helpful, especially for the AI-native generation. What I worry about is that the younger generation that is fully AI-native won't have the eyes to spot AI-generated content vs. human-written work anymore. Obviously, as technology advances, it will become more challenging. Still, if companies can disclose their model biases or even faults, then it would at least help people be more aware of potential misuse of information or have a sense of alertness or awareness of potential misinformation.

Do you have any ideas or recommendations on how we can help our next generation become more savvy about detecting AI and using it effectively?

I’m still struggling with it and trying to learn how to teach our next generation about this, especially as content becomes more “human-like”.

Anything else you’d like to share with our audience?

My goal is to provide insightful analyses on new technology, apply my knowledge of consumer tech strategy to frontier sectors, and use my bilingual/ bi-cultural background to bridge the gap for you, the audience. I want to meet more of you online and offline, build relationships and understanding, and maybe find some *spark* in unexpected opportunities.

There is so much noise out there. I aim to filter out the noise for you and provide thoughtful insights that could be valuable for your decision-making, whether you’re an investor, a researcher, a policymaker, or an entrepreneur.

While I can't predict the future, I work hard to analyze current developments and bring my connections and learnings across Asia. So, join my learning journey by subscribing to AI Proem!

See more on why I started AI Proem here.

Grace, thank you so much for joining me on this interview, and I wish you continued good luck with your newsletter and consulting business!

Interview References and Links

See Grace’s recent video interview with Alex Kantrowitz on Big Technology Podcast:

Grace Shao on LinkedIn

Grace Shao on Substack (AI Proem, Proem Communications)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! (One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊)
Share

📜 AISW #055: Daria Markava, Georgia-based high school student and AI enthusiast

Karen Smiley — Thu, 17 Apr 2025 10:06:05 GMT

Introduction - Markava

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Markava

I’m delighted to welcome Markava as my next guest for “AI, Software, and Wetware”. Daria, thank you so much for joining me for this interview! Please tell us about yourself, who you are, and what you do.

My name is Daria (she/her). I am originally Belarusian, but I’ve lived in Georgia for the last 3 years. I’m currently in grade 11 of high school.

I know school systems can vary worldwide - to confirm, does being in grade 11 mean that you have one more full year to complete before you graduate high school?

Yes, next academic year is my last year at school.

I noticed in your LinkedIn profile that you were a speaker in January at TEDx “Lisi Lake Youth”. Your topic was listed as “Artificial Intelligence: Youth Safety Matters”. Can you share a little bit about your talk?

Of course. The applications for the event came out around the time when the story of the boy who killed himself because of his obsession with a character.ai chatbot was on the news. I started reading more about what happened, and eventually I knew I wanted to speak about it, because it seemed like youth safety didn’t get as much attention in discussions on AI.

The main message I was trying to get across is that generative AI like that chatbot is very unpredictable, and that young people can’t always understand its risks and limitations when they encounter it online. It’s similar to social media, except the interactions with AI are so much more personalized, and that means you can be much more easily manipulated into interacting with it for longer (just a side note, I was talking mostly about AI chatbots for entertainment, like companions or social media bots, rather than the ones we use mostly for productivity).

The key issue here is that companies profit from those addictions, so they have an incentive to target such AI tools towards young people even though they must know how harmful it can be in the long term. It’s just very irresponsible.

I agree, and I’m so glad that you joined the voices speaking out about it!

(2) What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I have used AI quite a lot, mostly LLMs, and I’m also self-studying machine learning and deep learning. I don’t have any experience training my own models, but I’d say I have a pretty strong understanding of basic concepts. I also co-founded a club at my school that teaches primary / elementary students the basics of AI, and I’ve been leading it for seven months now.

That’s a very cool initiative, Daria! I would love to hear more. What prompted you to start the club? How many other people are in the club and teach with you? And how did you come up with the content you use to teach the elementary students about AI?

I had our STEM teacher approach me about the idea to start a new club about AI, as he’s heard that I’m interested in AI and have some experience. I decided to give it a go, and we came up with an initial outline of the curriculum. We started with very basic things, like what artificial intelligence even means, and then gradually moved on to more complicated concepts, like machine learning and neural networks. Usually each topic is related to the previous one, so most of the time we had no problem deciding what the next topic should be. Recently I’ve tried to put more emphasis on ethics, like data collection, privacy, bias, etc. We’ve used many online resources to demonstrate how AI tools work, like the teachable machine, but for theoretical content, I create the slides myself. We have about 10 students now.

That’s awesome.
(3) Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

I’m currently working on a research project on NLP; sometimes a paper I have to read has a lot of irrelevant information, so it would be extremely time-consuming to read the whole thing. What I do then is feed the paper into ChatGPT, and it gives me the specific points I’m looking for. After that I can ask for more details on certain points, or go directly to the sections of the paper that discuss these points.

That’s a good example. Do you have a sense for how often ChatGPT gives you accurate or inaccurate information? Like, when you’ve gone to a section of a paper to check a specific point, do you always find it, and does ChatGPT’s summary seem right?

The summaries are usually good enough, but they are only summaries. I haven’t encountered any misleading summaries or made-up points yet, but when I check the points in the paper, there’s always more to each one than what the summary mentions (which makes sense, because it’s just a summary). Overall, I think working with text documents is where ChatGPT performs the best for me, compared to other tasks. However, I still always check the original paper.

Got it. I’d love to hear more about your NLP research project! Is this kind of project common for students at your high school, or is this a special course? What is the goal of the project, and what tools are you using?

This is a special online program my school nominated me for, which is really exciting. It’s a research project, where you pick a question in the AI field and eventually write a 5000 word paper on it. My paper is on how NLP tools reinforce dominant cultural narratives. Initially I was planning to research machine translation and/or low resource languages, but as I was brainstorming ideas, they started leaning more and more into social and cultural sciences. I guess interdisciplinary studies and the intersection of technology and society interest me more than purely technical aspects.

My school offers something similar as an extracurricular subject, and I’m doing it too. For that, I research how AGI is portrayed in the media.

It’s great that you’re highlighting how AI can reinforce existing biases and looking at media portrayals. Can you share your papers here on Substack when they’re done? I’d love to read them!

That’s great to hear! I’d like to try to have them published if they turn out well - but if it’s allowed to post them without/before publishing, I’ll definitely share them.

I will look forward to that!
(4) If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

I usually avoid using AI for writing. There were multiple occasions where I needed to write a short text (I don’t remember the topics though), and I asked a few LLMs to write it for me. Every time the output text was too vague, too many phrases that served no purpose but to make the text sound sophisticated. Perhaps my prompts weren’t good enough, but in the end I decided that instead of prompt engineering, I might use that time to write the text myself. And just overall, I would probably never use AI to do the work for me. It can help with brainstorming, or streamlining routine tasks, but I always create the final piece of work myself.

I know a lot of writers who feel that way about using AI for writing, too - including me 🙂. Have you ever tried any of the generative AI tools for creating images, or songs, or videos? Or have you mostly avoided them, so far?

Yes, I’ve tried generating images and videos before. It was mostly for fun, I simply wanted to explore what these tools could do. Some of them were really good, but if you compare the generated outputs to the works of real human artists, they’re just not the same.

I tried using Sora the other day, and it generated a hilarious video of baking bread, with very poor object permanence or spatial understanding. I used it as an example of hallucinations in my AI club when we were studying generative AI, and the kids found it really funny how a hand had six fingers and a table suddenly turned into an oven in that video :)

That sounds like an effective teaching demonstration :)
(5) A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

Yes, the companies should definitely get the creator’s consent before using their work to train AI. Those people put so much effort into their craft, and for companies to just take that data and profit from it is simply unfair. Of course, compensating everyone whose data was used would be really challenging, but it doesn’t mean that stealing other people’s work is the solution.

There’s also an element of hypocrisy. When DeepSeek was released, I remember OpenAI accused the developers of using its GPT. So when a company uses artworks, books, etc, without the creators’ permission, it’s fine, but when another company (allegedly) uses a model without permission, it’s suddenly not okay to use someone else’s work to develop AI.

I think this is just an extension of a “hierarchy” that exists in society, where STEM and engineering are prestigious and respected, while creative work like art, literature, music, etc. is considered less important and not taken as seriously.

That’s a great insight on the hypocrisy of how some companies are handling questions of intellectual property.
(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

I definitely don’t think any companies providing AI are transparent about that. On the news I’ve seen many class lawsuits where writers sued AI companies for using their books to train AI without permission, for example the most recent one where Meta used books from LibGen. ChatGPT, which I use quite often, was also trained on the corpora from the web, and that could include copyrighted content. And of course, image and video generators probably used the most copyrighted works.

Yes, it’s definitely a problem. The last count I saw for generative AI lawsuits was 39 in the US alone.
I just found out last week, from the search tool in The Atlantic’s article, that at least 11 of my published research papers have been stolen and included in LibGen. That’s a tiny drop in a very big bucket, though. I know some writers who have had multiple full-length books stolen in LibGen, which is just awful for them.
If you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

I tried running a model (keras.io/examples/nlp/semantic_similarity_with_bert/) that determines semantic similarity between two input sentences. The data came from Stanford Natural Language Inference corpus (there’s an actual citation for it 1). The dataset basically contains pairs of sentences and their semantic similarity, and it was specifically created for training NLP models, which seems ethical.

It’s so great to see that researchers are making their data available in conjunction with their papers nowadays. And that they’re being transparent about how it was sourced or generated.
(7) As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

I haven’t heard about any cases from people I know, but one example I’d like to mention is a dataset that contains photos of people who died in prison. Obviously those people didn’t consent to their data being used in that way, and this example really shows how little regard AI companies have for our privacy. (To give credit where it’s due, I’ve learned about this case and many others from Dr. Joy Buolamwini’s book “Unmasking AI”).

Yes, that’s terrible. I know Dr. Joy’s book and the AJL website have some appalling examples of bias and exploitation that people should be more aware of - thank you for crediting her.
(8) Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs? How do you feel about how your info was handled?
Reference for this question: "Your Data is Worth Thousands—Here’s Who’s Selling It (And How to Stop It)"

I know that on LinkedIn and Substack, there is a “Block AI training” section in settings. It’s turned off by default, and I wasn’t even aware it was there in the first place until I decided to look through settings just to see what’s there. It’s good to know that there is still an option to opt out, but it’s presumptuous to opt everyone in by default without them agreeing. Even though you can technically opt out, the fact that your data and content could have been used to train AI before you discovered that setting is frustrating.

I completely agree - settings like “Block AI training” should not be OFF by default. I think the Substack setting is designed to block third party bots from using our content, not to control Substack using it. But still, the default should have been to block and not to allow.
LinkedIn definitely used our content before we had a chance to know about the setting, and the opt-out only applied to our future content. (And actually, there are two settings in different places that we needed to change to fully opt out, not just one. They were not up front about that, either.) I do know that people protected by GDPR were treated better with regard to AI training than those of us who aren’t protected. So that says it’s not a matter of not being able to do it, technically. It was a deliberate choice to exploit as many of us as they could legally get away with. Definitely not cool.

That’s a great point. They’re trying to grab as much data as they can before the owners realize it.

(9) Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Luckily hasn't happened to me.

Good to hear!
(10) Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

There are many things I can think of, but perhaps the biggest one is to start listening to experts, the people who warn about potential risks, rather than continue with the “move fast and break things” approach. A lot of educated, qualified people speak up about the harm AI can cause us, and ignoring them isn’t the way to earn public trust. So I think the best thing they can do is start thinking about people, and what’s best for humanity. I know it’s a bit unrealistic to expect this in our capitalist society, but it’s the only way to make sure that we benefit from AI as much as the companies promise.

Well said.
(11) Daria, thank you so much for sharing your thoughts in this interview. Is there anything else you’d like to share with our audience? And I’m curious, with everything you’re doing with AI, and everything we’ve talked about, I’m wondering: do you have a definite interest in pursuing a career in AI or in STEM? Or are you still mostly exploring at this point, as an 11th grader?

Thank you for this opportunity, Karen! Yes, I’ve already pretty much decided to pursue a higher education in AI and data science. I’m not sure yet if I’d like to have a career in industry or stay in academia, but I’ll probably figure it out during my undergraduate studies. One thing I know for sure is that either way, I want to center my work around responsible tech and ethical innovation.

I am so happy to hear that you’re aiming for a STEM career, Daria. You obviously have an aptitude for it, and we need more voices like yours shaping AI and other technologies in the future! Thank you again.

Interview References and Links

Daria Markava on LinkedIn

Daria Markava on Substack

Daria’s Jan. 2025 TEDxLisi Lake Youth talk on “Artificial Intelligence: Youth Safety Matters” (YouTube)

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that all content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊
Share

End Notes

Samuel R. Bowman, Gabor Angeli, Christopher Potts, and Christopher D. Manning. 2015. A large annotated corpus for learning natural language inference. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP). [pdf] [bib]

📜 AISW #052: Annie Ridout, UK-based freelance writer and creative entrepreneur

Karen Smiley — Thu, 27 Mar 2025 10:06:33 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - (she/her)

I’m delighted to welcome Annie Ridout from London, UK as my next guest for “AI, Software, and Wetware”. Annie, thank you so much for joining me on this interview! Please tell us about yourself, who you are, and what you do.

As a non-fiction author, I’ve had three books published: Raise your SQ (2023), Shy (2021), The Freelance Mum (2019). I’m also a freelance journalist and I’ve written for the Guardian, Red Magazine, Telegraph, Grazia and most other national newspapers and women’s magazines. And I write The Creative Way on Substack - all about living and working creatively.

Then, I teach everything I know about writing, business and creativity in online courses and through coaching/consultancy. So, that’s the career side of my life. But I’m also a mum to three children, aged 10, eight and five. Two are at school, one is homeschooled at the moment. Since learning about neurodivergence through my children, I’ve started writing about this, too - @raising_neurodivergence on Instagram and Raising Neurodivergence on Substack. It’s become my special interest. We live in London.

Annie, thank you for sharing that background. It’s inspiring to see that you’ve been so successful in your books, your newsletters, courses, shop, and other dimensions of your writing!
And it’s so interesting that you’re learning about neurodivergence through your kids, and writing about it. As you know, I work in AI, and one thing I learned as an AI leader is that neurodivergence is even more common among people who work in AI than in other fields. So I tried to learn as much as possible about how to build and support neurodiverse teams, and I wrote an article about it last summer.
One excellent article I just read this week here on Substack by was about the spectrum concept, and how labels like “low/high functioning” and “level 1-3” are ableist and not helpful. She recommends that we should talk about neurodivergent people not in terms of perceived shortcomings, but in positive terms, about what supports they respond well to. Like, “thrives with written instructions”. What are your thoughts on that?

I love that approach. I don’t really talk about the spectrum and I certainly never refer to someone as being on the spectrum, as this phrase is outdated and it has become derogatory. Though, there is a spectrum in terms of how an autistic person experiences life, but it’s a complex combination of factors, not just high or low. Similarly, ‘high and low functioning’ are also no longer used, as terms, as they’ve been replaced by ‘high/low support needs’ but there is also a question mark over those terms, as many neurodivergent people’s needs fluctuate.

My approach to raising neurodivergent children is to be open-minded, curious, to educate myself and to always remain hopeful.

Many of the posts I’ve been sharing on Instagram around neurodivergence are being viewed by hundreds of thousands, and sometimes millions of people. The message I am putting out is that our children bring something important to society. A different way of being, thinking, responding and communicating. When I say ‘different’, I mean different to the so-called neurotypical way of doing things.

There is a lot of support for this message, but I’m also being trolled more than I’ve ever been trolled before. There are people out there who are scathing about the positive message I want to spread, and it really highlights something deeply concerning about how neurodivergence is viewed by the general population. There is a lot of educating still to be done.

Getting back to AI, what is YOUR level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I don’t knowingly engage with AI and I never use it for writing. I’m pretty old-school when it comes to writing - I like a notepad and pen, and if I’m writing online, I don’t use AI to correct my spelling or grammar. In fact, I find it really annoying when it’s switched on because it miscorrects so often. That said, I know that AI feeds into most of our lives in subtle ways, so I’m certainly relying on it when it comes to the technology side of my business (which is all online, and largely based around social media). And in my private life.

I understand that! I’ve talked with a lot of people who want to keep their writing ‘by their own hand’. And even some people who are fans of AI tools have griped about auto-correct fails and annoying grammar checkers. Me, I’m one of those people who keeps auto-correct turned off on my phone 😉
Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

When my third baby was born, I got a Kindle for the first time. I’d always been into proper books but now, I knew I’d be lying in bed breastfeeding a lot and I wanted to be able to read in the dark. I bought the first novel that had been recommended and after that, the algorithm sent me so many brilliant book recommendations. I read about 10 books, back-to-back, and they were all so perfectly selected by, presumably, AI.

Thank you for sharing your experience with the Kindle and liking its book recommendations. So that would have been about 5 years ago - I guess before all of the AI-generated books began flooding Amazon recently ;) I’m wondering if you give Amazon active feedback on books you read - like, do you ever “Recommend this Book” or post book reviews on Amazon?

I often leave reviews on Amazon after reading books. As an author myself, I know how important this is, as lots of good reviews will lead to more book sales. This matters in financial terms, but also in respect of whether or not you get another book deal.

Do you still read books on Kindle, now that your kids are getting older? If so, do you find that Amazon’s ebook recommendations still hold up?

I still read novels on my Kindle at night. And then I read paperback non-fiction books in the morning, while I’m drinking my coffee. The recommendations aren’t as good as they used to be, for some reason. And so now, I am working my way through the Women’s Prize for Fiction longlist.

I don’t have many gadgets in my house but I have a TV and I subscribe to Netflix, so again, I rely on AI helping me to choose the next drama to watch. I also subscribe to Spotify and love the playlists that are recommended to me. And when I’m out and about, I use Google Maps to find the best route to walk or drive.

Netflix, Spotify, and Google apps are all good examples of AI in our everyday lives. I’m curious: one thing I’ve read and heard from musicians is that AI-generated songs have been crowding out music by humans on Spotify. Do you know if any of the songs that have been recommended to you may have been AI-generated? Or do you generally know the musicians Spotify is recommending for your playlist?

I always know the musicians I’m listening to. Often, I find them on Instagram first via a live performance that has been filmed. But my kids like some really shit music that is probably AI-generated, about people farting and similar. It hadn’t actually occurred to me that it’s AI-generated.

That’s funny 😆 If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

As I mentioned above, when I’m writing, I avoid anything that will get between me, my mind and the words that roll out. I’m definitely inspired by content I find online, that will often be showing up because of AI (on Instagram, for instance), but AI doesn’t get to directly touch the way my thoughts turn ideas into a piece of writing.

I fully support that. By the way, if you haven’t yet seen the “Created with Human Intelligence” badge initiative that started, you may want to check it out. (The one I use is at the end of this interview.)
I’m wondering if you have heard about the ‘Human Authored Certification’ announcement from the Authors Guild? They are now supporting author members to self-certify that AI was not used for anything except “minimal, trivial uses” in a book. We had a Notes thread on this and one commenter, , mentioned that Writers Guild of Great Britain (WGGB) is a similar organization in the UK. What are your thoughts about that kind of self-certification by authors? Would seeing that self-certification by an author be meaningful to you as a book reader? Would you as a writer want to self-certify your next book?

I don’t know that I would add that badge to my writing. And I’m not sure that it would make a difference to me, as a reader. I assume that a novel I’m reading has been written by the author. If they’ve used AI for spelling and grammar, I don’t need to know that. Or, to be honest, if they’ve used it for help with the plot. But I do want to know that the author is a real person and similarly to the music I listen to, I do tend to know who the author is before I read the book.

Got it. A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get Consent from (and Credit & Compensate) people whose data they want to use for training? (the “3Cs Rule”)

I don’t mind AI being trained in this way, but I do think it should be done with consent. My publishers asked if I would give permission for AI to be trained on my first two books, for a lump sum, and I agreed. I feel confident in my ever-evolving writing style and ability to innovate with words. Also, the fact that I am a real person, with real-life experience, still matters to my community.

I’m happy to hear that you were given a true choice about whether AI was trained on your books, and that you were compensated.
Do you know if that agreement covered use of your words only, or did it also include the audiobook versions (was it your voice, or someone else’s)?

I don’t actually know about the audio version. I narrated my second book, someone else narrated the first one.

(6) As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

No, I don’t think we are told where these ideas come from. I use Canva to design posts for Instagram and when I think about it now, perhaps that website has been trained on AI, pulling ideas from designers.

Yes, Canva is one of the design tool providers that uses AI. When I first started digging deeply into AI ethics last year, Canva looked like one of the more ethical design tool providers. And when there was that big uproar about Adobe’s ethics last summer, I looked into Affinity by Serif as an affordable alternative and put them on my ‘ethical shoestring’ list - they looked pretty good. But it came out that Canva acquired them and acquired Leonardo AI in July 2024, and some questions came up then about their ethics and plans. So I didn’t pursue buying an Affinity license. I haven’t looked into Canva’s ethics since then.
I noticed that most of your newsletter posts have your own drawings - which are cool! - or what look like stock images. Do you ever use Canva for making post images for your Substack newsletter? Or have you tried the built-in Substack options for adding free stock images from Unsplash, or generating an image?

I draw my own images for The Creative Way Substack and don’t use any stock images. For the Raising Neurodivergence Substack, I used a post template that my sister, a designer, created for me. But occasionally, I’ll use a stock image - of a pile of books, for instance - and change it to the brand colours.

That’s a great drawing! Thank you for sharing it here.
As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

No.

Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML?

No.

Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact.

A lot of us were angry with Meta for using our content to train AI, without getting our proper consent. I deleted thousands of my posts, in protest.

Did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

No.

Yeah, I never did join Instagram because I didn’t like their privacy policies, but I did purge my personal Facebook account last summer. I deleted just about everything I’d put there over the past 15+ years. It was my only real option when they refused to honor my opt-out request. I know a lot of people who just left FB completely. But I know some who feel they can’t leave because they’ve established a business presence there. It sounds like Instagram is like that for you, maybe?

I recently deleted my Facebook completely (not for the first time - I did this about 15 years ago, too). Twitter (or ‘X’), too. But yes, Instagram is part of my business. I’ve grown a community there under my own name, and now I have a very fast-growing community around the words I’m sharing on neurodivergence. I’m not leaving the platform just yet, but I’m not afraid to have some time away, or pause awhile.

Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Not that I know of.

That’s good to hear! So your phone number and email address haven’t been sold to data brokers, you haven’t gotten spam or junk calls or texts?

Ha - yes, I’ve definitely experienced spam calls, emails and texts. But it doesn’t happen very often and I’m not sure who sold that information.

Yeah, sometimes it’s really hard to tell where those people get our data.
Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

They can keep reminding people that there’s space for AI and for the real human experience. It doesn’t need to be one or the other. When I hear people saying that AI will take over and the human race will no longer exist, I roll my eyes. I think this is a really pessimistic attitude and that we can take a more optimistic view on it all.

I agree - I think the extreme views on both ends, hype and doom, are unrealistic and not helpful. I try to encourage people to think of AI as ‘Augmented Intelligence’ or ‘Automated Intelligence’ that complements human intelligence. So, like you said, not one or the other.
Anything else you’d like to share with our audience?

I mentioned my books and Substacks at the start. I also have an Instagram account under my own name, @annieridout, where I share poetry and poetic passages about… the human experience. The engagement has dropped a lot recently, since I deleted all those posts, but I don’t regret it. We need to exercise our power and freedom within these technologies, sometimes.

The Creative Way (with Annie Ridout)

Let's talk about Instagram.

A few years ago, one of my friends asked another of our friends whether Instagram was ‘dead’. I felt a bit panicked, listening in, as much of my work stemmed from Instagram. Journalism commissions, book deals, online course sales…

a year ago · 174 likes · 11 comments · Annie Ridout

Agreed! Annie, thank you so much for sharing your experiences and time for this interview. I’ll embed the links to your Substacks and your IG here so that people know where to find you. I wish you continued success in all of your writing!

Interview References and Links

Annie Ridout’s website

Annie Ridout’s Instagrams: neurodivergence, personal (poetry)

Annie Ridout on LinkedIn

’s newsletters on Substack:

Sample posts from The Creative Way:

The Creative Way (with Annie Ridout)

It's not procrastination, it's LIFE

I have this new approach to starting the week. Instead of waving goodbye to my children and rushing to my desk to action all the ideas that have landed over the weekend, I pack a rucksack and head to the gym…

a year ago · 86 likes · 7 comments · Annie Ridout

The Creative Way (with Annie Ridout)

I lost 31 paid subscribers ($2130) in December

When I first joined Substack in March 2023, I’m not sure I believed I would actually be able to earn a living from it. After all, I’d spent years publishing my writing for free on various platforms, alongside traditional journalism and writing books…

a year ago · 1000 likes · 20 comments · Annie Ridout

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that all content in these interviews will be human-created:

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊
Share

📜 AISW #044: James Presbitero, Philippines-based writer and strategist (AI, Software, & Wetware interview)

Karen Smiley — Mon, 03 Feb 2025 11:06:49 GMT

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Photo of James Presbitero - provided by James and used with his permission.

Interview -

I’m delighted to welcome James Presbitero from the Philippines as our next guest for “AI, Software, and Wetware”. James, thank you so much for joining me today! Please tell us about yourself, who you are, and what you do.

No, I should be thanking you! Thank you for this opportunity, Karen. I’m James Presbitero, a content strategist and writer from the Philippines. I’m the owner of Write10x, a fast-growing publication centered around the promise of helping writers be 10x faster, better, and more human. Currently, I talk a lot about my philosophy regarding AI and writing, which I call the “Mindful AI Mindset”. Its main idea is that we should use AI mindfully because mindless use of AI exposes us to danger now and in the future. I’m seeing how that philosophy applies to writing.

Thank you for that introduction, James! We’ll definitely include a link to your Write10x newsletter for those of us who care about AI and writing. For now, can you elaborate a bit on what you see as the main dangers of AI for writing, and how it can be used mindfully?

Right, I’ve explored this topic in depth across several articles, and I feel like there are layers to the dangers of AI in writing.

On an individual level, the main danger is over-reliance on AI. When AI tools first started gaining traction, many people jumped on the hype, treating AI as a shortcut to writing success or a quick way to make money. While there’s more awareness around mindful AI use now, I still see people relying on it without developing core writing skills.

Here’s the key thing to remember: AI and writing are two very different skills. AI can only scale what already exists. If you lack foundational writing ability, AI will simply amplify that lack—it won’t create a compelling, authentic voice for you. For new writers especially, learning AI before mastering the craft of writing can be dangerous. It creates a risk of over-reliance, and over time, it can actually erode your writing skills if not used thoughtfully.

On a broader scale, the biggest danger of AI in writing is its potential misuse by harmful actors. I’ve written about scenarios like AI-powered influencers—essentially AI posing as real people to spread propaganda or manipulate audiences. This could be exploited by political parties or unethical companies, like Ponzi schemes, to generate content at scale and lend credibility to their actions.

As AI becomes more advanced, I believe the unethical use of it will proliferate unless we act now. We’re looking at a significant challenge in the future if frameworks and safeguards aren’t put in place.

That’s why I advocate for democratizing AI knowledge. The more people understand AI and how to use it responsibly, the better equipped we’ll be to counter harmful uses. By giving AI tools to individuals and organizations with ethical intentions, we can create content that challenges and outpaces bad actors.

In practice, this looks like integrating AI into the writing process rather than letting AI be the process. Developers and users must take accountability for the impacts of their AI systems. It also means being aware of the ethical repercussions of the content we create and being transparent about how we use AI in our work.

Ultimately, mindful AI use involves blending human creativity with AI’s efficiency—keeping humans, their values, and their intentions at the center of everything we produce.

What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I use AI a lot, both for my work and personal writing practice. Primarily, I use ChatGPT. I use it professionally, as a superpowered writing assistant. It helps me strategize, brainstorm, write content, and create image assets. I also use it personally for much the same ways.

Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

Oh I’ve used ChatGPT for all levels of writing and research, from ideation down to editing. At first, I’d say that ChatGPT wasn’t giving me satisfactory results at all. But as the technology improved and I learned to use it better, the rate at which it gives me favorable results has only increased. That's true for both creating text content and image assets.

Can you share some specific examples of text or image tasks that you used ChatGPT for? Maybe one task where it didn’t give you satisfactory results, and one where it did?

AI has been incredibly reliable for generating content ideas, especially when they’re tied to something specific like a particular offer or a broad topic. In fact, I’d say AI is better than me at producing linear, structured content ideas—like generating a content pillar or brainstorming related topics around a central theme.

However, I find that AI struggles with providing ideas that feel truly original. That said, it’s excellent at helping me harness and refine my own original thinking. For example, if I have a unique concept that hasn’t been explored much, I can use AI to brainstorm angles, identify potential gaps, or even predict what my audience might be thinking. It acts as a valuable sounding board for amplifying my own creativity, but for broader, more generic ideas, AI often outpaces me in efficiency.

On the other hand, one area where I’ve had less success with AI is image generation. I primarily use ChatGPT because it meets about 90% of my needs, but when I’ve experimented with tools like DALL-E for image content, the results haven’t always been satisfactory. I suspect this is partly because I’m still learning how to craft optimized prompts for image-based tasks. Image generation feels more complex and finicky, so it’s an area where I’m still figuring things out.

Aside from ChatGPT, I also regularly use Grammarly for grammar suggestions, though I'd say I only take its suggestions about 80% of the time.

I did some experiments with Grammarly last year when I was looking for a tool to measure readability of my writing. I also found I disagreed with some of its suggestions ;)
If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Yes, there are use cases where I deliberately avoid using AI. The biggest one would be using AI as a friend or therapist. A lot of AI tools are marketed as companions—they can mimic a specific character you choose or act like someone straight out of a movie. But something about that doesn’t sit right with me.

Talking to AI for leisure, especially about personal vulnerabilities, feels wrong. Maybe it’s because, deep down, I know they don’t truly understand me. They’re just pre-programmed to generate responses that are most likely to feel favorable or seem appropriate in the moment. They don’t empathize, they don’t understand, and relying on them in such an emotional capacity doesn’t feel genuine.

Marketing AI as a “companion” also opens a whole can of worms that we need to approach carefully. Misuse in this area could lead to serious consequences. For example, there’s the recent situation with Character AI, where their model allegedly encouraged a vulnerable child to engage in antisocial behavior, and ultimately resulting in an unfortunate death. That’s exactly the kind of issue that arises when we don’t adopt mindful and ethical practices around AI companions.

Those are great points, James. I interviewed a psychotherapist last year, before that Character.AI incident, and asked her about the use of AI for ‘therapy’ or ‘friendship’. Her take was that while she could see a few situations where people could find those kinds of tools supportive, that they really weren’t likely to help most people solve their loneliness or their problems. And those tools really need to be designed and trained with safety at top of mind, just like robots should be. It sounds like it fell tragically short.

I completely agree—there’s still a lot of work to be done in terms of creating frameworks and guardrails for using AI as a companion.

Funny story, though: my girlfriend actually uses AI as a companion sometimes. She describes it as a kind of interactive journaling—like journaling, but with a journal that responds back. She’s mentioned that she enjoys the impartial responses and the brief acknowledgment of her feelings that AI can provide. It’s not like talking to a person, but she finds it helpful for small moments of self-reflection.

I also have a friend who’s used Character.AI before, primarily for entertainment. She often talks about the weird, wacky responses the AI gives, and while she doesn’t take it seriously, she finds it fun to use.

Both of them, however, are very aware that these tools are just that—tools. They know they’re talking to chatbots, not real people. This recognition has its benefits, like appreciating the AI’s impartiality and its almost textbook-perfect bedside manner. But I can imagine how dangerous these tools could become for people in vulnerable positions—those who might rely on AI companions as a crutch for loneliness or deeper emotional struggles.

On a more troubling note, I think there’s an even larger danger here, one that Scott Galloway introduced me to: the potential for AI companion tools to be weaponized by bad actors. For instance, corporations, terrorist organizations, or even foreign adversaries could use these tools to radicalize lonely individuals over time. These could be people in positions of influence, like government workers, or simply individuals with vulnerabilities that can be exploited. By creating AI companions that are endlessly patient, kind, and subtly manipulative, bad actors could gradually steer these people toward harmful ideologies or actions.

We already know that loneliness is at an all-time high, and covert tactics like this have existed in other forms, such as espionage. There’s no reason to think bad actors wouldn’t take advantage of this incredibly powerful technology. That’s why I believe it’s crucial to implement robust guardrails for AI companion tools—to minimize these risks and ensure that their use is safe, ethical, and supportive rather than harmful.

A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get consent from (and compensate) people whose data they want to use for training?

Yes, I think it’s ideal for all AI companies to disclose where they’re getting their data from and to compensate people or parties whose data they use, depending on the type of data. Transparency and fairness should always be prioritized.

That said -- and this might be a hot take -- but I’m not particularly concerned about where companies are getting their data—unless, of course, they are violating laws or infringing on copyright. Right now, this issue is still being decided by the courts, as they work to determine whether AI data usage constitutes a copyright violation or if it falls under an extension of fair use.

In my own view, the way AI models use data feels more like an extension of fair use. They’re not copying anything outright—they’re using large datasets for research and pattern recognition. It’s akin to how a person might read a wide range of books, internalize them, and then be able to reference or write in the style of those books.

Because this approach isn’t a direct reproduction, I think it’s challenging to establish a legal precedent that would outright punish it. After all, most creative endeavors draw inspiration from existing works in some way. However, I’m leaving it to the courts to make that final call, as these are complex issues that go beyond personal opinion.

Well, we have well over 30 active lawsuits in the US alone on AI infringement and ‘fair use’ - it’s definitely going to take a while for those to get sorted out!
As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Oh, I don’t think they’ve been transparent—at least not in a way that’s obvious to most users. Or maybe it’s just that I haven’t been particularly focused on it myself. That said, whenever I use an AI tool, I don’t see the sources of their data being disclosed upfront. Maybe the information is available if you dig for it, but only people who are really interested or knowledgeable would even think to look.

I think AI tool providers could do a much better job of publicizing where their data comes from. For example, they could include this information in their marketing materials or prominently display it during the signup process. It needs to be visible and accessible—not buried in fine print—so people are aware without having to go out of their way to find it.

As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

I don’t have a specific case to share, but I feel like this has been happening for a long time. It’s just getting more attention now because of tools like generative AI. Big companies like Google, Facebook, and Microsoft have already been using personal data for years—to draw conclusions about users, send targeted ads, or design marketing campaigns that appeal to specific demographics.

Even older, non-generative AI systems have used data in some capacity. The difference now is that people are more aware of it because of how visible and interactive generative AI tools are.

That’s an astute observation - use of our data is certainly not new to AI or to generative AI in particular.

My current mentality when using AI tools is to assume that any data I willingly input—like my writing style in tools like ChatGPT or basic information during signups—could be used by the system. To me, that’s just the nature of using technology today. I do not know if this is a healthy outlook, but I feel as if I’ve accepted it as a fair tradeoff for the capabilities I enjoy. Of course, it’s a different matter if this data is accessed without consent, like retrieving private information I didn’t knowingly share.

I believe that we should push for better regulations and policies around tech, and AI specifically. But law by nature moves very slowly. It will never catch up to tech’s exponential growth. Therefore, the onus is on us, as users, to be mindful of the information we put online. We should also take steps to protect data we want to keep private. It’s all about practicing due diligence in an increasingly connected and data-driven world.

Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact.

I know for sure that big tech companies like Google, Meta, and others use my data for training purposes. But I’m generally comfortable with the level of data I expose to these companies.

One thing that did surprise me recently was LinkedIn. A few months back, there was news about LinkedIn training its own AI algorithms using content posted on the platform. While I’m not entirely sure if that initiative went through, it definitely caught me off guard. I don’t feel LinkedIn was particularly transparent about this practice or emphasized how user-generated content might be used to train their AI.

Yes, there was a huge uproar about LinkedIn opting many of us in by default, without our consent. Our only ‘choice’ was to opt out going forward. In the US, we didn’t have any recourse on them using our older data. But in EU countries covered by GDPR, they were protected from being opted in. Were people in the Philippines also opted in without consent?

The Philippines has a GDPR equivalent, called the Data Privacy Act. It mandates that companies obtain explicit consent from users before processing their personal information. So users should actively agree before their data is used, often by ticking a box or similar action. Implied consent, where users are automatically included unless they opt out, is not recognized under Philippine law.

But I don’t think LinkedIn specifically asked me to opt in for that, or if it did I forgot about it already. I should probably be more mindful about that.

In general, it feels like companies don’t provide real, meaningful choices when it comes to opting out of these policies. Often, the changes are buried in T&Cs, which most people don’t read. Even if you do read them, declining usually means giving up access to the platform altogether—a decision that isn’t practical for most people, especially professionals who rely on services like LinkedIn for their careers.

While I personally haven’t felt violated in how my data has been used, I think companies should be more upfront about their practices and make opting out easier and more accessible. Transparency and consent should be prioritized, and users should have control over how their data is used.

Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Not really—not for me personally. I have a personal philosophy that what makes you human, and even what makes you an excellent writer, isn’t just your words or writing style. It’s your ideas, your core values, the way you interact with people, and the things that AI inherently cannot recreate.

That said, I do recognize that this problem exists for many people. For example, I know a number of writers whose livelihoods have been impacted by AI. They’ve been accused of producing AI-generated content even when they’ve heavily edited their work—or in some cases, when they haven’t used AI at all.

Yes, that’s a downside of the ‘arms race’ in AI - where people are developing AI-based tools to detect AI-generated content, and those tools have false positives which unfairly flag human-created content. Did your writer friends have any recourse to appeal when their work was flagged as AI-generated? Did this block them from publishing, or from earning their living?

No, often they didn’t have any recourse. Especially when the “correction” happened after the contract had been signed. This has affected their earnings, mostly because of the additional wasted time they took to “fix” whatever issue may have been found. However, I’ve seen some other writers put a limit specifically for AI edits. I don’t do freelance clients, but I would recommend that. They should also stipulate in their contract their stance on AI-based tools to filter out clients who might be very nitpicky about things like that.

I’ve also seen how AI is affecting graphic artists and designers. AI tools are increasingly encroaching on their industry, automating tasks that once required their unique skills. This has caused significant challenges for many creatives.

I think more steps need to be taken to alleviate the kind of displacement AI is creating across industries. Free training programs, for instance, or better education on how roles might evolve due to AI could help affected professionals transition more smoothly. These efforts could make a big difference in minimizing the negative impact of AI while empowering people to adapt to new realities.

Those are good suggestions and I agree that as a society, we need to look out for the people affected by these seismic changes.
What are your thoughts on how shrinking the pool of human creators may dry up the flow of fresh, original content which AI tools need to continue to improve?

I'm not sure that's true.

I see two key points regarding this concern:

1. The creator pool isn't shrinking; it's expanding.

We're living in a golden age of the creator economy. It has never been easier to become a creator thanks to widespread access to education, tools, and training. The barriers to entry are lower than ever, and this means more people are stepping into creative roles.

With such a massive influx of creators, there’s an abundance of fresh ideas being added to the ecosystem. AI benefits from this growing pool of human creativity rather than being limited by it.

2. AI thrives on good content, not necessarily human-exclusive content.

AI doesn't need content to be "purely human-made" to improve. It needs content that is thoughtful, high-quality, and meaningful. When creators use AI tools mindfully to amplify their skills, they produce better work, which raises the quality of content available for AI to learn from. Over time, this creates a positive feedback loop where both AI and creators grow.

However, there’s a flipside.

If AI tools are misused to churn out low-quality or inauthentic content, the quality of the ecosystem may deteriorate. This is why promoting mindful use of AI is crucial—it ensures that AI and creators work together to elevate the baseline quality of online content.

In the end, AI can be a tool to expand humanity's creative boundaries, democratize creativity, and push innovation further. But how society handles AI—encouraging responsible use while preventing misuse—will determine whether it enhances creativity or diminishes it.

That’s a thoughtful reflection on why mindful use of AI is so important - thank you, James. The main concern I hear is that so much AI use at present is not mindful, and that low-quality ‘slop’ being created with AI isn’t adding value for society, and isn’t useful for the AI systems to use for further training. It’s hard to measure ‘quality’ systematically, though!
This has been a great conversation! My last question: 🙂
Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

I think the most important thing AI companies need to do to earn our trust is to be fully transparent about what they’re doing and how they’re doing it. One of the best steps they can take is to actively contribute to the creation of policy guidelines and laws for ethical AI use—not just for today, but also with a focus on future implications.

On a more practical, day-to-day level, I think it would be beneficial for companies to establish some kind of educational arm. For example, they could create blogs or YouTube channels that explain in detail the capabilities, potential harms, and dangers of their AI products. This would demonstrate that they’re genuinely committed to keeping humans at the center of their processes, rather than just focusing on profits or technological progress.

By educating the public and being transparent, they can show that they value accountability and the long-term well-being of their users. That kind of effort would go a long way in building trust.

That’s a great point. I do see some startups and initiatives that focus on educating the public about capabilities and risks. But the companies that are creating the tech don’t seem to be doing that - they seem to be avoiding it like the plague.

You're absolutely right—many tech companies seem reluctant to educate the public about AI’s risks and capabilities, likely because it could slow their growth or raise regulatory scrutiny.

That said, this creates a paradox. To pressure these companies to act responsibly, more people need to be educated about AI to build enough social and political leverage. But waiting for profit-driven companies to take the lead on education is unrealistic.

While the responsibility shouldn’t fall entirely on users, we do play a part. By using AI thoughtfully and sharing knowledge, we can create a ripple effect. The more people understand AI, the better positioned they’ll be to demand the policies, regulations, and transparency needed to ensure safer AI use.

It’s a cycle—educating others leads to greater awareness, which then builds the pressure necessary for meaningful change. The key is starting small but staying consistent.

Anything else you’d like to share with our audience?

Thank you so much for this opportunity, Karen! I’ve really enjoyed answering these questions and articulating my thoughts on such a significant topic as AI.

For anyone interested in the intersection of AI and writing, and how to use AI mindfully, I’d love to invite you to subscribe to Write10x. It’s a newsletter packed with highly actionable tips, prompts, templates, and guidelines that you can immediately apply as soon as you read them. The goal of Write10x is simple: to help writers, content creators, and businesses write 10 times faster, better, and more authentically human.

As a subscriber, you’ll also receive some amazing freebies, including The Outline + Prompt Kit and the Active Prompt Vault—both designed to supercharge your content creation process.

And if you’re interested in scaling your content strategy authentically with AI but aren’t sure where to start, feel free to reach out. I’d be happy to chat about how we can make it happen together!

Once again, thank you for having me, Karen. All the best to you and your audience!

Thank you for your time and for sharing your thoughts on AI, James!

I’m truly honored to be here, Karen.

Interview References and Links

LinkedIn: linkedin.com/in/jamespresbitero2022/

Medium: medium.com/@jamespresbiterojr

Substack: substack.com/@jamespresbitero

’s “Write10x” newsletter:

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊
Share

📜 AISW #040: Aysu Keçeci, Turkey-based sustainability consultant

Karen Smiley — Mon, 20 Jan 2025 17:02:28 GMT

Introduction - Aysu Keçeci

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Aysu Keçeci

I’m delighted to welcome of from Turkey as our next guest for “AI, Software, and Wetware”. Aysu, thank you so much for collaborating with me on this interview! Please tell us about yourself, who you are, and what you do.

I’m a 24 years-old passionate sustainability professional with a diverse background. I’ve experiences with Sustainability Business Development, AI & Sustainability, Corporate Sustainability, and Entrepreneurship. I’m currently living in Istanbul, Turkey, and preparing to move to Europe in the upcoming months.

My journey in sustainability began four years ago when my project, WE—a gamification-powered application designed to encourage recycling behavior, integrated with an IoT smart recycling bin—was recognized as one of the Global Top 50 projects in the Google Solution Challenge, an international sustainability competition. My story was also exclusively published on Google's official blog.

Since then, I’ve constantly worked to expand my skills and create impact.

Following this achievement, I initially founded the startup WE, offering solutions for corporate environments and partnering with the region's top industrial corporations. However, when the materials and chip crisis arose during the pandemic, I shifted my focus from hardware to software.

Then I founded GateZero — a gamification-powered SaaS platform designed for corporate sustainability. The platform leveraged AR [Augmented Reality] and gamification to promote sustainable behaviors among employees, tracking their progress through interactive sustainability dashboards, and provided automated reports for the ESG [Environmental, Social, & Governance] reporting process of the companies.

In addition to my entrepreneurial journey, I’ve also worked in sustainability-focused roles in technology companies. I’m currently taking consultant roles in the companies mostly for their Sustainability Business Development needs, but also providing content creation support around sustainability, AI and green innovation.

I also run Road to Earth 2.0, a blog dedicated to sustainability, AI, and systemic change. It’s a space for curious minds seeking to better understand our world, the systems we live in, and the future we’re shaping.

The blog is a place to envision a feasible alternative world with reimagined building blocks. It draws on elements of history, sociology, economics, and philosophy to offer a more holistic, macro overview perspective—aiming to contribute to the number, awareness, and, hopefully, the vision of those striving for this change.

That’s a commendable mission!
What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology?

I’ve come to realize that AI, ML, and analytics have seamlessly integrated into nearly every aspect of our lives—far beyond the moments we consciously notice. I’ve actively used these technologies in my previous startup experiences and in the products we developed. For example, in GateZero, I utilized AI for generating reports for companies, designing tools that analyzed current states, and provided actionable insights.

When it comes to generative AI, I frequently use it for media content creation. I’ve experimented with many different tools to generate visual content, as the models are constantly being updated. However, after certain updates, I don’t always get the results I’m looking for, which keeps me in a constant cycle of discovering new tools. It’s fascinating to see how much progress has been made in just a year—it’s truly impressive.

Additionally, since English is not my native language, I often use AI for quick grammar checks, which has been incredibly helpful.

One tool that has become a constant part of my daily life recently is Rewind. I feel like I’m just at the beginning of this journey, but the idea of having my augmented brain partner—an AI companion of sorts—within my digital world is exciting and full of potential.

Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

One specific example that comes to mind is when I was able to make website adjustments or add features to an application we were developing, even without much coding knowledge, by simply following GPT’s guidance. It was incredible not just to depend on my developer teammate but also to contribute directly to the project itself. In a way, AI feels like it’s reducing the exclusivity of knowledge, and I think that’s a beautiful thing. It eliminates the need to spend time on mechanical tasks that don’t require creativity, allowing more room for innovation.

Another instance was when we created a promotional video for GateZero, where I heavily relied on AI tools. AI proved invaluable in terms of cost-effectiveness and speed, enabling me to produce a video that met my expectations without being an expert in video production. This was about a year ago, and considering how much AI has advanced since then, I’m sure even better results could be achieved today.

I used several tools, including Runway, Eleven Labs, Adobe AI tools, and D-ID, to bring this project to life. With Runway, I generated unique stock videos and replaced my own appearance with a fun virtual character to better showcase the AR product idea—one of my favorite uses of AI. For voice generation in parts of the video where I wasn’t speaking, I used Eleven Labs, and when I ran into synchronization issues between visuals and audio, I resolved them using D-ID, avoiding the need for re-recording. Additionally, I used Adobe’s AI tools for tasks like resizing mockups and enhancing generated visuals.

Those are great examples of how you combined AI tools to accelerate your work. Can you talk a bit about your experiences with using AI in GateZero operations? Specifically, “I utilized AI for generating reports for companies, designing tools that analyzed current states, and provided actionable insights.” Which AI tool did you use? What parts worked well, and what parts didn’t work so well? For instance, did you ever find mistakes or ‘hallucinations’ in the reports or bugs in the tools, or did you need to iterate on the insights to get something you could deliver to companies?

Using the GPT-3 API, we developed several tools. For instance, one tool provided recommendations to improve a company’s sustainability performance based on their existing sustainability reports, structured within a framework we designed. Another tool analyzed sustainable behaviors among employees and generated actionable outputs and improvement suggestions.

While we didn’t encounter any critical errors—mainly because the tools weren’t handling highly sensitive tasks—there were instances where the model presented outdated information, as GPT’s knowledge didn’t include the most recent years at the time.

If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

When it comes to AI—or technology in general—I consider myself an early adopter. I enjoy experimenting with and using new tools, and I rarely feel hesitant about it. At most, a few seconds of worst-case misuse scenarios might cross my mind before diving in. But that’s just part of any technological advancement, and ultimately, how we choose to use these tools—positively or negatively—is in our hands. For this reason, I actively use AI and love trying out new things.

That said, I try to distance myself from the culture of fast consumption as much as possible. When I want to deeply research or understand something, I still turn to books or human-written content rather than relying on quick AI-generated answers. These sources provide depth and nuance that AI often cannot replicate.

For example, in the context of art, while I find AI-generated music assets impressive and occasionally enjoy listening to them, I still prefer music created by human artists. One of the things that makes me connect with art is the story behind it and the bond I feel with the artist.

That’s a sentiment I hear often (and agree with) about music and art!
A common and growing concern nowadays is where AI and ML systems get the data and content they train on. They often use data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get consent from (and compensate) people whose data they want to use for training?

This doesn’t personally worry me too much, as long as something unique to me isn’t being copied or commercialized without my consent—like training an AI on Scarlett Johansson’s voice and monetizing it, for example. I think my lack of fear stems from my belief in the necessity and potential of AI.

I believe ethical AI regulations, restrictions, and oversight are absolutely necessary—and urgent.

I know many people feel that they’re helping humanity by generously allowing their work to be used for these purposes without compensation. That is commendable, and it sounds like you feel that way. Where I get uneasy is with companies basically forcing *everyone* to do this, against their best interests, when a few other people will profit greatly from it. And across the world, not everyone can benefit from the tools. That doesn’t seem fair.

I’m also unsure how realistic it was until today to expect a clear and enforceable structure for distinguishing between public and protected data on the internet, especially considering the massive scale of information involved. AI models like LLMs require immense amounts of data to train, and naturally, they rely on public data from the internet, the world’s largest information source.

In an ideal world, there would be clear distinctions about which data is shared under specific licenses and permissions, ensuring only appropriately authorized data is used. However, in today’s reality, the internet lacks such a structured system to enforce these distinctions effectively, making it no surprise that issues arise when working with data at this scale.

Even the multi-billion-dollar film industry, supported by thousands of pages of legal texts, laws, and teams of lawyers to protect its content, struggles to prevent pirated copies of films from being shared as public content. Once these works are circulated across the internet in this way, they are falsely treated as public domain, regardless of their original protected status. Without a centralized authority or reliable tracking mechanism to determine content ownership and associated rights, it becomes nearly impossible to manage and monitor millions of pieces of data in such massive data pools.

Now, consider the vast amounts of data on the internet, much of which was previously deemed to hold little or no value. It’s hard to imagine this data existing in a structured format, clearly labeled as public or protected, and being managed effectively across the digital landscape.

I think things will work differently from now on, especially regarding consent, data ownership, and the compensation of contributors, as we have only recently encountered and faced the practical implications of these concepts.

I believe much of the concern comes from fears about certain creative professions becoming obsolete. However, when it comes to creativity, I’m not convinced it can truly be replicated or surpassed by AI. Creativity, at its core, is unique and holds an intrinsic value that can’t be replaced. Of course, this depends on both the creator and the audience, but as long as people continue to seek genuine creativity—and I believe they will, despite commercial pressures—its value will endure.

Studies are already showing that AI tools ARE cutting into livelihoods for many creators - writers, artists, musicians, and others - and diminishing their value in real terms. My sense is that the concerns go beyond obsoleting creative work, though, into fundamental questions of fairness and ethics.

Until AI advancements, humanity and nature have been somewhat isolated in their relationship, and when it comes to tackling climate change or global challenges, we’ve struggled to make sufficient progress on our own. We could certainly use all the help we can get. Now, we’re developing technology that can actively participate in problem-solving and design processes, even pointing out, “No, this isn’t beneficial”, and producing solutions that may surpass what we could imagine unaided. In that sense, AI holds immense promise.

One major concern is that most of the business people who say they want to solve all of the world’s problems are profiting tremendously from the tools they build with the data they are taking without compensating people, so their work isn’t exactly altruistic.
As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

Unfortunately, there is a clear lack of transparency when it comes to the data sources used for AI models, especially with something as massive as large language models (LLMs). Even the companies developing these models often don’t have a full understanding of where all the data originates.

In my experience, when working on products I’ve developed, we made a point of tracking and sharing the data we used and its sources because it was an important aspect for the company. However, since the scale of these products wasn’t as vast as an LLM, it was much easier and more practical to implement.

That makes sense.
As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Since it’s something I encounter so frequently, I think I’ve become much less attentive to it. One of the more innocent AI interactions that sometimes feels strange is when WhatsApp suggests the next word I’m about to type—and gets it surprisingly accurate.

As I mentioned, because I encounter these situations so often, I’ve likely become desensitized to them, which makes it harder to recall specific examples. However, the moments that stand out the most usually involve audio rather than text. For instance, after discussing a particular topic with a friend, I often end up seeing ads or solutions related to that conversation shortly afterward.

That’s a common experience - several of my past interview guests have shared stories about this, too!
Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs? How do you feel about how your info was handled?

Rather than examples of companies that might use my data, it’s easier for me to think of examples of those that likely don’t—like Apple, which feels more secure to me. Many companies probably ask for consent or disclose my data usage, but it’s often buried within pages of terms and conditions or casually mentioned during account creation. Keeping track of or preventing such usage feels like such a high-friction process that I usually don’t manage to do it effectively.

From my experience working directly with users, I’ve seen that almost no one reads the agreements or notices shown at the start of an app. In daily life, we’re exposed to so many of these interactions, often alongside regulations, that it creates a sense of fatigue—I think many of us feel this way. These regulations sometimes feel less like meaningful protections and more like a way to confirm that they’ve obtained our consent, even if it’s not entirely clear or deliberate. As a result, I rarely feel like I’m given a real choice in these matters.

Yes, there was a study that over 90% of people never read the terms and conditions. And since they’re so hard to understand, it’s hard to fault them for it!

At least for the cases I can identify, I choose not to give permission. However, when I consider the data I’ve unknowingly shared or how it may be inferred, those efforts don’t seem very effective in providing actual protection.

On occasion, particularly with beta versions of products I like, I do allow my data to be used if I believe it could genuinely benefit the product’s development.

Unfortunately, I’m no longer surprised when I find out a company has used my data. If anything, I react more strongly when I see that it has been used in a beneficial way. In many ways, this issue isn’t new with AI—it’s part of the broader, ongoing problem of data security that remains unresolved.

That’s true. You mentioned Apple. They have definitely cultivated a reputation for protecting privacy. So a lot of people are disillusioned by the recent revelations about Apple settling the lawsuit over using Siri on the iPhone to listen to people’s conversations and then selling the data about the conversations. How do you feel about learning that?

Every software and system we use collects data for analytics, telemetry, or to analyze our preferences and digital behaviors. At this point, what matters to me as a user is whether that data is collected in an anonymized way or in a manner that can be linked to my digital identity. This distinction makes all the difference in how I perceive the ethics and privacy implications of such practices.

That’s a good observation. In cases where the data is being sold to brokers and merchants for them to target selling to us, the data IS connected to us as individuals. If it’s being aggregated and used for general product improvement, like you mentioned with a beta version, that’s more acceptable. We don’t always know how they will use it, though.
Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

While not directly related to AI usage, I recently experienced a challenging situation involving data insecurity. In Turkey, there’s an electronic platform called E-gov, a government service that provides citizens with access to various digital services. It enables many administrative processes to be handled efficiently online. However, your inclusion in this system is, of course, independent of your consent or choice.

Not long ago, it was reported that the personal data of nearly all citizens was likely stolen through this platform and is being sold online on platforms like Telegram for very small amounts of money. The stolen data includes identity information, addresses, phone numbers, information about relatives, and more.

In my case, someone who had only my phone number was able to find my address and use it to threaten me, claiming they could harm my relatives. This was a deeply unsettling experience and a stark reminder of how data insecurity can have real, tangible consequences.

I must emphasize again: data insecurity isn’t something new that arrived with AI. In fact, it has likely followed a steep upward curve since the advent of the virtual environment.

Oh wow. Having all of the government’s data on people breached is a nightmare! How awful that you were threatened like that 🙁And you’re right, data theft and exploitation started long before generative AI tools became so broadly used.
Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

I don’t think this is a problem we don’t know how to solve—it’s something that companies either choose not to do or, in some cases, are unable to do. Transparency, radical transparency, and efforts to close the gaps that allow for misuse are what we all expect from these companies.

We also need information from companies like the MAANG giants or OpenAI that shows they’re trying to understand and address what’s happening within their systems. Yet, even they admit in interviews that they’re often uncertain about certain aspects. Sometimes, they avoid sharing information simply because it doesn’t reflect well on them.

For example, just as they fail to fully disclose the emissions data from the data centers driving the AI boom, they are equally opaque about the implications of their systems. What’s different here is that they often don’t fully understand the potential consequences themselves—or they don’t seem particularly motivated to figure it out, given the cutthroat competition of this capitalist environment.

I agree with you that the biggest obstacle is the financial factors that steer company decisions. Until they have a financial incentive to earn our trust by operating transparently and ethically, and by doing diligence on understanding the risks and issues, most of them won’t bother.
Anything else you’d like to share with our audience?

I’d like to invite everyone to join me in imagining Earth 2.0—a feasible alternative world with reimagined building blocks.

When we consider that our systems operate on supply and demand, the reason we haven’t solved the climate crisis or other global problems isn’t because we lack the means but because we haven’t truly wanted to solve them. So why, despite our constant complaints about the state of the world, haven’t we taken real action? Perhaps it’s because we struggle to imagine the rigid building blocks of an alternative world.

What if we knew there was a tangible alternative worth fighting for? Would things change? That thought is what inspired me to start my blog.

If you have ideas about what Earth 2.0 could look like—with all its foundational systems—I warmly invite you to share them. Let’s reimagine, learn more, and explore the possibilities and pathways together.

And Karen, thank you so much for inviting me. It was such a pleasure meeting you and doing this together!

Thank you for sharing your experiences and thoughts, Aysu! And I love that you’re focusing on inspiring people to envision what our new world could and should be. Good luck, and I hope we can continue this conversation and explore the parallels between the AI ecosystem and the climate crisis and other aspects of sustainability. Motivating companies and people to care about solving both is a core challenge!

Interview References and Links

Aysu Keçeci on LinkedIn

on Aysu Keçeci’s Substack:

Leave a comment

About this interview series and newsletter

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I Don’t Use AI”:

Series Credits and References

If you enjoyed this interview, my guest and I would love to have your support via a heart, share, restack, or Note! One-time tips or voluntary donations via paid subscription are always welcome and appreciated, too 😊
Share

📜 AISW #037: Arun Mozhi Maruthamuthu, India-based senior DevOps engineer

Karen Smiley — Thu, 02 Jan 2025 11:06:42 GMT

Introduction - Arun Mozhi Maruthamuthu

This post is part of our AI6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and content.

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

🗣️ Quick note for those who prefer listening to reading: Substack now supports read-aloud natively in the Substack app. If you try it on this post, let me know what you think!

Photo of Arun Mozhi Maruthamuthu - provided by Arun and used with his permission

Interview - Arun Mozhi Maruthamuthu

I’m delighted to welcome Arun Mozhi Maruthamuthu from India as our next guest for “AI, Software, and Wetware”. Arun, thank you so much for joining me on this interview! Please tell us about yourself, who you are, and what you do.

I’m Arun Mozhi Maruthamuthu, with over 13 years of experience in the IT industry, and 8+ years specifically in cloud infrastructure. In my current role as a “Member of Technical Staff” at Wind River Systems, I focus on a variety of key areas such as DevOps, Site Reliability Engineering (SRE), managing cloud platforms and product owner responsibilities.

Throughout my career, I’ve had the opportunity to work across a diverse range of technologies, from traditional IT systems to cutting-edge cloud-based solutions. My expertise has grown from foundational IT roles to leading cloud infrastructure projects, DevOps initiatives, and SRE practices, allowing me to drive efficiency and scalability in the systems I work on.

On a personal note, I am always excited to have discussions with you, Karen. Whether it’s during our one-on-one sessions, internal meetings, or even discussing AI-related topics on LinkedIn, I value these conversations as they provide both learning and the opportunity to exchange ideas. I look forward to continuing to collaborate and explore new ideas together!

I’m so glad you agreed to this interview, Arun - I always enjoy talking with you as well 🙂
What is your level of experience with AI, ML, and analytics? Have you used it professionally or personally, or studied the technology? (or built tools using the technology, …?)

In my previous organization, I worked on ML Proof of Concept (PoC) projects. Currently, I am part of Wind River, where I access AI-related products for both internal use and my own learning. To deepen my understanding, I am taking basic ML courses through various learning platforms and plan to explore data analytics in the future. My experience so far spans the professional and personal use of AI.

Can you share a specific story on how you have used a tool that included AI or ML features? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

WhatsApp AI recommendation : AI suggests quick, context-based replies to my messages. It saves time in formulating long messages. It also gives a wide range of inputs to various personal fronts too.

Drawback: It gives only text based suggestions. Images and document related features are not available.

How often are the suggested replies useful to you?

I would say that this feature provides streamlined and summarized information compared to what Google typically offers. It simplifies searches by giving direct answers, which is especially helpful when you need quick insights. However, there are times when it provides overly long or detailed responses, which can feel less user-friendly.

When AI-powered chat features were initially launched by Meta, I experimented with chatting as if it were a friend. I even vented about some personal things out of curiosity, and I was surprised to see it respond with friendly and empathetic replies. It gave me a glimpse of how AI could evolve into a conversational companion beyond its practical uses.

Currently, WhatsApp’s AI chat feature is text-based, and voice-based interactions are not available. If WhatsApp enables voice-based messaging for AI chats in the future, it could significantly enhance the user experience.

With voice-based AI, users could interact more naturally, allowing them to send queries or tasks through voice commands and receive responses audibly or in text form. This would be especially helpful for hands-free use, such as while driving or multitasking. It could also make the platform more accessible for individuals who prefer verbal communication over typing or those with disabilities.

Combining voice capabilities with AI-driven personalization would make WhatsApp a truly versatile tool—not just for chatting but also for managing daily tasks, resolving queries, or even having friendly, interactive conversations. This evolution could further solidify WhatsApp as a leader in both communication and AI integration.

Tagging in Facebook: I think I first noticed AI on Facebook a long time ago when I saw how it could recognize people in photos and suggest their names for tagging. I explored a little more and found that a “Deep Learning Algorithm” was used which recognizes unique facial features.

How often are the tagging suggestions correct, and how often are they wrong?

I don’t recall any instance where Facebook’s auto tagging feature gave an incorrect suggestion; it almost always worked perfectly for me. However, I’ve noticed that this feature is no longer available in my account. It seems Facebook has disabled the auto tagging feature in India, likely due to privacy concerns and regulatory compliance.

Oh, that’s interesting to hear!

ChatGPT: Once I started using ChatGPT, I found myself relying less on Google for debugging and coding. It’s incredibly helpful for quickly getting started, even in areas outside my expertise.

Drawback: You do need to exercise caution, as it sometimes provides multiple suggestions, some of which might not work as expected.

Can you share a specific example of a code suggestion that didn’t work as you expected? Do you find that ChatGPT’s code suggestions work better on some programming languages than others?

I can’t recall the exact scenario, but I faced issues with Python package installations on macOS. The commands provided weren’t working as expected, likely due to administrative access requirements. While ChatGPT suggested a temporary workaround of using a virtual environment for installations, it didn’t provide a permanent solution to address the underlying admin access issues. This limitation made troubleshooting and resolving the problem more challenging.

Alexa: Alexa has been a lifesaver for managing daily tasks. It helps me stay on top of bills, manage my shopping list, and enjoy music tailored to my preferences. One of its biggest benefits is reducing screen time for my child. Instead of spending hours on screens, my kid uses Alexa to play music for dancing and as a tool for studying, asking questions, and exploring new topics. It’s a versatile assistant that makes life more organized, entertaining, and educational for the whole family.

Those are great examples of uses of a smart home assistant. May I ask, how old is your child?

My son is 6 years old and currently in Grade 1. At his school, they start using an iPad lab from Grade 3, and I’m certain they incorporate AI features into their learning tools. For now, they use the Macmillan Education app for English lessons. The app offers a range of resources like video-based stories, grammar practice, pronunciation support, and extra exercises and worksheets, which I find quite beneficial for enhancing his learning experience.

My son is also learning Abacus at a private global institute called “BrainoBrain.” They conduct international-level online competitions where AI plays a significant role in evaluating tests and ensuring fairness through fraud detection mechanisms.

I can go on with much more features in my phone like auto-spamming messages/calls, photo memories, Auto-correction and Text Prediction, Face ID, camera enhancements etc.

AI is becoming an integral part of our daily lives, and as it evolves, there may come a point where it’s hard to distinguish between actions performed by humans or AI. It’s clear that AI is increasingly shaping our world and our future.

That is definitely true - it’s almost impossible to avoid AI in our everyday lives nowadays!
If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

Since AI learns from the data we input, I’m always cautious about using it for official work and avoid sharing any confidential information.

That’s wise.

I’ve had a few regrets with Google Maps, especially in situations where it led me down the wrong path or didn’t provide the most effective route. One time, it directed me onto a narrow mud road where my car couldn’t go any further, causing a significant detour. Because of such experiences, I now only rely on Google Maps when I’m traveling to unfamiliar areas, and I double-check routes when possible to avoid any surprises.

Ha, that’s happened to us a few times, too! Once, we were in California, and we decided to check out the town of Mountain View (Google’s HQ). Google Maps routed us to some odd location that was nowhere near Mountain View! My husband still refers to that incident as an example of why we can’t trust it 😆

Absolutely, reliance on AI tools like Google Maps has become a significant part of our daily lives. It sometimes lead to mishaps/accidents☹️. Here are two recent incidents from India that highlight such scenarios:

Accident Due to Incorrect Navigation: Link
In a recent case, Bareilly, Uttar Pradesh, where two men lost their lives after following Google Maps' directions. The app mistakenly directed their car onto an incomplete bridge over the Ramganga River, causing the vehicle to plunge into the water.
Dangerous Routes Suggested in Remote Areas: Link
In another incident, travelers in a remote region were misled by Google Maps onto an unsafe, isolated path due to incorrect or incomplete mapping. The detour caused them significant delay and put them in a risky situation.

There have been a few times when I tried creating images of specific objects with backgrounds, but the results didn’t turn out as expected.

Yeah, image tools can produce strange results. This past summer, my friend Stella tried to use an AI-based image tool to create a promotional image for her book (on an older woman writing a romance novel), and the results were SO odd - the tool put a white beard on the woman! (link)

Oh God!! I think we can use the current AI image generation tools only for entertainment purposes to create unrealistic images such as your friends 😆, definitely not for professional use.

Additionally, we don't allow our child to use Alexa or any voice recognition tools without supervision. We’re cautious because he might ask questions without fully understanding the consequences or implications of the responses. It’s important to ensure he uses these tools safely and responsibly.

That is super important and it’s smart that your family is being so careful about avoiding unsupervised use by your child.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get consent from (and compensate) people whose data they want to use for training?

Ethical AI companies should be transparent about how they collect, use, and store data. Clear, accessible consent forms should explain the purpose behind data collection, how it will be used for training AI, and who will have access to it.

If AI tool companies profited mainly because of using one person’s data (non publicly available), they should give compensation in my point of view.

I agree, Arun. I also think the 3Cs (consent, credit, and compensation) shouldn’t be limited only to content that’s not ‘publicly available’. Some AI company execs have used ‘publicly available’ as if it’s the same as ‘public domain’, and it’s not. As one example, YouTube videos are publicly available, but they’re generally copyrighted and not public domain. AI companies shouldn’t be making money off them without giving the creators the 3Cs for their work.

Yes. I completely agree with you on this point.

As a user of AI-based tools, do you feel like the tool providers have been transparent about sharing where the data used for the AI models came from, and whether the original creators of the data consented to its use?

As a user I feel, the transparency regarding the data used to train AI models is often insufficient. While some companies may be more transparent than others, many don't clearly communicate where the data comes from, whether consent was obtained, or if creators are compensated. Ethical AI development should prioritize transparency, clear communication, and fair compensation to build trust and accountability.

Yeah, most AI companies today are not transparent about data sourcing. I like to call out those few who are, whenever I hear about them; I think they deserve our attention and support!
As consumers and members of the public, our personal data or content has probably been used by an AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Social media content is often leveraged by AI tools in various ways. For example, Facebook's face recognition technology has historically used personal photos without explicit consent to suggest tags.

Similarly, targeted advertisements rely on AI-driven data sharing across platforms. If I search for a product on Amazon or Google, I immediately see ads for that product on social media. It’s frustrating and concerning to feel that my privacy has essentially been sold.

Definitely. And an even bigger concern I hear about the photos is: it’s one thing to opt in for letting them use the photos to suggest tags. It’s another thing for Facebook or Instagram or another service to use those photos for all kinds of other purposes without our consent.
The privacy concerns around ads are important too. I’m always looking for ways to turn off or disable that data sharing, or to keep my browsing private or separate, to minimize that.
Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact.

While AI companies provide legal disclosures, the language is often hidden in long terms and conditions. At least I don't read terms and conditions in all the cases, who will read very long terms and conditions. They are playing cleverly to make us fools and we are falling for the same.

Ideally, companies should offer more accessible, user-friendly explanations and clear options for opting out of data use for AI training.

Absolutely. Yes, most people don’t read the Terms and Conditions - there was a recent study that said 91% never do. It’s hard to fault people for not reading them when they aren’t understandable anyway!

Do you feel like you have a real choice about opting out, or about declining the changed T&Cs?

I was not even aware of this opt-out option itself. Most apps can't be used at all, if I decline T&Cs. So there was no choice left out.

How do you feel about how your info was handled?

I dislike my personal information being used without my consent. For example, when I share my mobile number during purchases—whether online or while using a debit or credit card—it often gets misused. I receive spam calls for promotions, donations, or credit card offers, sometimes even from banks where I don’t have an account. These unsolicited calls are frustrating and disruptive.

I agree - the sale of our data and this kind of spam are so annoying.

After hearing multiple instances, I deleted Facebook and Instagram long back for safety concerns.

Good for you!
Has a company’s use of your personal data and content created any specific issues for you, such as privacy, phishing, or loss of income? If so, can you give an example?

Both privacy and phishing concerns: Recently, I received repeated calls claiming to be from FedEx about a returned parcel that I had no knowledge of. When I finally inquired, they said the parcel contained illegal items and was addressed to a location in Kolkata, with my mobile number linked to it. I've never been to Kolkata in my life. The caller insisted I travel to Kolkata immediately to file a complaint with the cyber security department or, alternatively, click a link they provided to file it online. I told them I would look into it and ended the call immediately.

Wow.

This incident highlights the rise in phishing attempts. On top of that, I often receive spam messages, get added to various job-related groups, and encounter scams offering pay for writing product reviews or posting advertisements. These constant intrusions make it difficult to protect personal privacy and avoid potential scams.

It is definitely hard to stay vigilant about these risks.
Public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

The most crucial step AI companies must take to earn and maintain public trust is ensuring transparency. Users need a clear understanding of how their data is collected, utilized, and protected, as well as the societal impacts of AI technologies.

One practical solution is to provide users with a comprehensive overview of all the platforms where their data is being used. Companies should offer an easy-to-use system—perhaps linked to an email address or phone number—allowing users to access a complete list of services and systems utilizing their data. Additionally, users should have the ability to opt out and request the deletion of their data across all platforms seamlessly.

I agree, Arun - this is where we need to go, as a global society. The big challenge, I think, is for us as consumers to make it enough of a business concern for companies that they will invest in doing it.
Anything else you’d like to share with our audience?

Thank you so much, Karen, for this opportunity. Through this process, I’ve come to realize the significant impact AI has on our daily lives. Data privacy has become a key concern for all users, and if this issue is properly addressed, AI can truly make our lives easier.

Thank you for taking the time for this interview and sharing your experiences, Arun!

Interview References and Links

Arun Mozhi Maruthumuthu on LinkedIn

Leave a comment

About this interview series and newsletter

This post is part of our AI6P interview series on “AI, Software, and Wetware”, launched in August 2024. It showcases how real people around the world are using their wetware (brains and human intelligence) with AI-based software tools or being affected by AI.

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I don’t use AI”:

We want to hear from a diverse pool of people worldwide in a variety of roles. (No technical expertise in AI is required.) If you’re interested in being a featured interview guest, anonymous or with credit, please check our guest FAQ and get in touch!

6 'P's in AI Pods is a 100% reader-supported publication. All new posts are FREE to read (and listen to). To automatically receive new 6P posts and support our work, consider becoming a subscriber (it’s free)!

Subscribe now

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

If you enjoyed this interview, I’d love to have your support via a heart, share, restack, Note, one-time tip, or voluntary donation via paid subscription!
Share

📜 AISW #036: Anonymous8, USA-based automotive business / presales leader (AI, Software, & Wetware interview)

Karen Smiley — Thu, 26 Dec 2024 11:06:45 GMT

Introduction - Anonymous8

This post is part of our 6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and content.

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Photo by Michal Czyz on Unsplash

🗣️ Quick note for those who prefer listening to reading: Substack now supports read-aloud natively in the Substack app. If you try it on this post, let me know what you think!

Interview - Anonymous8

Today I welcome my 8th anonymous guest, from the USA, to “AI, Software, and Wetware”. Thank you for making time for doing this interview with me! Please tell us all about yourself and who you are and what you do.

Fantastic. Karen, thank you so much for, number 1, just inviting me to be part of this platform and share some of my thoughts, insights, and opinions about this type of technology and how it's transforming many different industries, but also how it's supporting or enabling other types of technology to reach its full potential. So, thank you for that.

My professional journey consists of diverse roles, predominantly focusing on business strategy and planning. I have been fortunate to work at three global OEMs [Original Equipment Manufacturers] and two Tier 1s [major suppliers to OEMs]. My responsibilities varied from strategic planning, strategic alliances, product marketing, product planning, and engineering - fuel economy optimization, software development to testing and validating the crashworthiness of vehicles.

In the early stages of my career, I was directly involved in engineering work, specializing in electronics and software development. Taking control models and turning them into code and establishing the feature test cases to ensure the robustness and readiness of the vehicle to withstand adverse road conditions, such as ice or snow.

I subsequently joined a software solutions company, where I focused on establishing strategic partnerships and solution development, driving value creation for our customers. My career has come full circle as I transitioned into business development and sales.

It's definitely an active and interesting cycle that we're in right now, with all the movements around different types of AI and machine learning. Can you tell us about your level of experience with AI and machine learning and analytics, and how you've used it professionally or personally, or if you've ever studied the technology?

I may no longer be directly involved in engineering, but indirectly, I'm very much immersed in it through the products I sell. At my last two companies, we've been leveraging AI components to enhance user experiences. In the automotive industry, consumers now expect smartphone-like responses, with the ability to download new apps or services instantly, or to receive frequent technology updates. This necessitates a level of "smartness" to deliver personalized experiences.

We're using AI to not only collect data, but also to ask the right questions and derive insights that drive meaningful actions and enhance the user experience. Vehicles are no longer static devices, but dynamic, living machines that can be upgraded, updated, and modified based on user needs and preferences.

Professionally, it's fascinating to witness the evolution of technology and its capabilities. On a personal level, one of my career goals is to serve on a board. In preparing for this, I've been using AI tools to refine my bios and resumes, asking questions like, "Is there a more effective way to present my experiences?" Sometimes, the AI output isn't quite right, but with different queries, I can mix and match to craft my message.

You’re using the AI tools with your wetware, not instead of it. 🙂

I've heard people speak in a fear-like manner about AI and how it's going to basically take over and take jobs and blah blah blah. And I'm like, AI is at this state where it's still in its infancy. Though it's proliferating in popularity, it still can't stand on its own. It's not like it's going to take over today, and it's not going to take over in 10 years. But what I see it as right now is an augmentation to human capability. And because I see it as an augmentation to the human capability, that means we still have to apply some of our intellect to identify and decipher, “Okay, this is what it spit out, but how do I make this coherent in a way that can convey the message and the depth that I want it to?”

So that's kind of how I've used AI professionally and personally.

Okay, thanks!

Oh, and you asked, have I taken any classes? I haven't taken any classes, but one thing I definitely want to do is an AI course for business applications. I would like to understand “How can I, today, better leverage some of the tooling that's out there?” Because it's already available, though like I said, it's in its infancy. But there's still quite a bit of opportunity where I could, in small pilots, test things out just to see what it comes up with.

I interviewed somebody last week who’s working in the area of AI marketing ethics (). And he's developing a course for people to learn how to use AI ethically to support them in their marketing activities. There's probably a course out there for what you're looking for.
I'm curious to know: in your business development and presales role, obviously there's a lot of communication. Have you used, or resisted using, tools like ChatGPT to help you with authoring those communications? Or are you really only using it for your board member materials?

I primarily use AI tools for personal purposes, as my customer interactions typically revolve around our products and internal content, where I don't currently require AI assistance. However, I am open to exploring AI tools for professional use in the future.

Currently, my professional use of AI is mainly through smart features in Outlook that correct grammar and punctuation. While these features have been helpful, I haven't yet plunged into more extensive AI usage for full-blown communications.

You probably spend a lot of your time in meetings.

Yes!

Zoom and Teams have different AI-based features. They have automatic captioning, and they have meeting summarization and features like that. Do you use any of those features on a day-to-day basis?

While I haven't yet implemented this approach, I'm open to considering its potential applications. Although I haven't yet explored extensive AI usage in my professional communications, it's not outside the realm of possibility. For instance, AI-powered transcription of recorded customer or partner meetings could prove beneficial in capturing important details for future review or reference.

I'm always curious, when people talk about meeting summarizations and communications, about confidentiality. Especially if you're talking with partners or customers, confidentiality can be a big concern in using those tools. And allowing the tools to record and transcribe and summarize can be a risk. Whatever confidential information you may be discussing, using the tool could expose that info to the people who make the tool. And from there, the info can bleed out into the wild, right? So some people that I've talked to have actually hesitated to use those tools - not because they don't like the tools, but because customer or client confidentiality, or even doctor-patient confidentiality in health care applications, becomes a factor.

Indeed, my hesitancy in adopting AI extensively in business stems from concerns about clarity on acceptable boundaries. With the technology evolving rapidly, it's crucial to understand where the lines are drawn to avoid potential risks such as leaks that could lead to competitive disadvantages or detriment to the business, partners, or customers. Given the fluidity of the situation, it's challenging to know precisely where these boundaries lie. It feels like we're building the vehicle while simultaneously driving it, which makes it difficult to confidently navigate the landscape.

Right. One of the analogies I've been using with folks is that they worry about “AI is so complicated, and I don't understand it. I don't know how to build a machine learning model.” But you shouldn't have to know how to be a mechanic in order to drive a car safely. We just want people to learn how to ‘drive’ AI safely and securely and not have to worry about their information being compromised or something leaking out. So they don't have to know how to build a model.
But we do want to encourage and support people in learning how to ask good questions about, “Well, okay, if I use this tool, what data are you going to save and where is it going to go and who's going to have access to it and how can I control it and see what you're doing?” Learning to ask those good questions is sort of an AI literacy or data literacy aspect that I think is important.

Absolutely. I completely agree.

You mentioned one specific story about using an AI tool, ChatGPT or Copilot or something like that, to help you with your board materials, your bio, and your resume.

Exactly.

I’m curious about any other specific examples you’ve seen. Maybe one of the features that your teams are developing, where they've tried using AI to make something smarter? For instance, being smarter about software updates or some specific feature. Usually it takes some iterations to get AI models right - some things work quickly, some things need refinements, and biases often need attention. Do you have any stories about times where you've tried either building features, or using features in somebody else's tool, where it didn't work well and you had to adapt to that?

Just similar to what I was sharing about the fact that certain sentence structure or something, sometimes it didn't convey the meaning that I was trying to get across. And so that's where human intelligence comes in. Like, this isn't exactly what I was aiming for. And then you have to read, just like you would do if you were training a model, you have to ask it more questions to get to where you're trying to go.

In terms of a specific example where something didn't go as planned, I would say maybe Insta Headshots. It's an AI-generated photo shop, if you will. And it will take your picture in these natural scenes. And sometimes it looks pretty accurate - like, it's pretty good. It'll color your suit in different colors. Have you heard of it?

No, I need to look that up. (instaheadshots.com)

Sometimes you just gotta play with these things, right? Because they're not going to get it right. But it's kind of cool that such a thing exists, right? Where you can take a picture and it's a natural picture and you're like, wow.

I saw a post recently on Bluesky or LinkedIn where someone had posted about having computer-generated headshots and a headshot taken by a professional photographer. And could people tell which one was which, which one did they like? It's interesting to see the style.
Image generators are one of the AI tools that people talk a lot about. And some people have tried the music generation tools. But one of the big concerns there, and this is kind of coming up as a general theme, especially with the generative AI tools, is: Where did they get the data and the content that they use to train their tools on?

Exactly.

Sometimes they take data that users put in online systems or they published online. And companies don't always say what they're going to do with our data when we sign up for it. Some are taking copyrighted content or content that, yeah, it's publicly available, but that doesn't mean it's free to use for commercial purposes. So there's a lot of concern around that not being ethical.

Indeed, public skepticism surrounding AI tools and their handling of data is a major concern. Ensuring data privacy and creating trust between businesses and consumers is vital. One way to achieve this is by focusing on transparency in privacy policies, granting users control over their data, and maintaining clarity in every interaction where consumer data might be involved.

The transparency should extend to businesses seeking permission to use consumer data for their purposes, creating a clear line of communication between the consumer and the company. This level of transparency is crucial in an age where smartphones and apps frequently request access to personal information, such as location or cookies.

My friend's startup aims to restore trust between businesses and consumers by emphasizing data privacy and transparency, which I find fascinating. The work they are doing is a crucial building block in addressing the concerns and building confidence in the responsible use of consumer data.

Well, there's consent, and then there's informed consent, where people actually understand what they're consenting to. You see a lot of these terms and conditions on websites and apps, and they're 10 or 20 pages long, and they're in legalese that even lawyers struggle to understand.
So the prompt might say, “We need to use your pictures to generate this head shot for you”. And you say “Yes”. You're allowing them to do that. But then maybe they're also taking those pictures and using them to train models that they're going to sell to somebody else. And you don't really want to consent to that, but that wasn't the question that you were asked.

Exactly. Yeah.

I really want to hear more about your friend's startup. I want to look into that.

It's called “360 of me”. Actually, I'll send you a link. (360ofme.com) I would love to introduce you to Kara Grasso. She is absolutely phenomenal.

I would love to meet her and talk with her about the whole idea of how to rebuild the trust between people and the AI and tech companies. We're seeing a lot of data out there now that trust has been diminishing recently. Because people are starting to realize, “You're doing what with my data?” Or “You're capturing this or you're using pictures of me or my family for what purposes?”
And there was that big uproar over the summer about Meta and what they're doing with people's content, and Suno stealing all the YouTube videos. Trust has been severely damaged in the whole global ecosystem, I think. Someone who's actually looking at what companies need to do to rebuild that, I would love to see what they're doing and what her thoughts are.

Oh, yes. I'll definitely introduce you. She has lots of thoughts. And, actually, Kara is the Chief Revenue Officer, but the CEO of this company, Cindy Warner, oh my goodness. This woman is a powerhouse, just incredibly brilliant. I'll send you some videos, and actually you could probably just YouTube her. You'll find the videos.

She is on a mission to restore trust, in terms of consumers being clear on how their data is being used. And not only how their data is being used, but also, she wants to restore the ability for the consumer as a contributor of the data to get a piece of the pie in terms of money, right? And so she's also on that campaign too. And I was like, you know, you're absolutely right. It's the user's data that's being used to train these models. Why shouldn't they get a cut of whatever revenue you're generating?

Yeah. In the big picture, what I've been hearing people call it this year is “the 3 C's” for creative rights or contributors rights, and it's for Consent and Credit and Compensation. In other words, if someone uses content from a book that you wrote, and they cite it, you should be credited for it.

Exactly. It's like royalties. It's royalties in music. If you're going to use my stuff, then I should get a cut.

Yes. Exactly. Looking for the 3 C's. Some people say 4 C's, with Control as the 4th C. But Control is really, I think, part of Consent.

Yes. I agree that you can use it, and I get to say how you can use it.

Yes. That's it. So 360ofMe sounds like a neat company; thank you for sharing that. It's really cool.

Oh, absolutely. Yeah.

We talked about consent, and you mentioned transparency. Do you know of any companies that have been transparent in a good way about the data for tools that you've used? Or more generally, do you feel like companies have been transparent about where they got their data and what they've done with it?

Despite the existence of privacy documents and terms and conditions, the tools I have used still lack complete clarity regarding data usage. These documents, often spanning multiple pages, fail to provide comprehensive information on how my data might be employed, potentially requiring legal consultation to fully understand.

This lack of transparency persists even when using platforms or tools that claim data isolation or containment within an account. While it may be stated that the data will be used for the company's purposes, the specifics of how and what data will be utilized remain unclear, leaving users uncertain about the potential implications.

You mentioned the features for making cars smarter. With automotive, there's been a lot of discussion about the privacy (or lack thereof) for data, for people's data that their cars are gathering. Cars have so many sensors. And now they have interior sensing cameras, so you've got images of people too.
There's some legal protection for privacy for content on our phones. But if you connect that phone to your car's infotainment system, suddenly, they can get data and content out of our infotainment systems without a warrant. There are really powerful things that we can do with the data from the car, but it also creates some serious privacy risks. So there are some tradeoffs there that we need to look at, and I'm really curious to hear your thoughts on this.

Let me reframe your question. You want to hear my thoughts on how we deliver these advanced features that people want for their cars and still protect their privacy? With the data that's being collected from them while they're in and around their cars?

Yes.

I think this is part of what 360ofMe and the platform that they've built out, right, is -assessing and understanding that. Because once you buy the car, it's yours effectively, but at the same time it has this feedback loop back to, let's call it the mother ship, so that way it can continue to be this dynamic machine throughout the life of that vehicle.

To establish transparency, it is crucial for users to be aware upfront of the ongoing data collection from various sensors and services within the vehicle. This data is gathered by the OEM, but to enhance the vehicle's functionality and improve the user experience over time. By clearly communicating the purpose and benefits of data collection, users can feel more confident that their information is being used responsibly and to their advantage.

Some of the concerns come in with features like, for instance, some of the OnStar systems. They're always keeping track of where somebody is. And that's a potential privacy concern. Then there's information about how somebody drives, and the different insurance companies. You mentioned partnerships and strategy. It's a whole ecosystem, right? Where the OEMs have data from the car that could be used for purposes that maybe the person didn't intend or want.

Yes. Exactly. And this is the beauty of a company likes 360ofMe. They will facilitate, let's call it, that transaction to make sure it's clear to the user, the end user of that product, and also, to the business customer, making sure they have all the necessary rights and privileges to obtain, process, and utilize that data.

Yeah. We could probably talk for a long time about this! It’s such an interesting area, and I'd love to hear more sometime about these kinds of features. Do you have an example of a smart feature that you're trying to build? And where would the data come from for that?

I mean, we don't even have to think about the future. We could think about even in the past, like Elon Musk, right? When they developed the Tesla, I think it was 2008-2009, he set up a subscription service where you could enhance your range, based on your driving behavior. And so, based on your driving behavior that got set up, then you could identify if you needed to subscribe for more range. And then they would do a software update based on you subscribing. But at least you had a way to gauge if you were overextending yourself and needed more range.

Okay, that's a good example of how a smart feature could be utilized.

In the future, infotainment systems in vehicles will become more dynamic, allowing users to add and download various apps and services for convenience, as well as safety features as new regulations come into effect. This will be facilitated by connectivity and software updates, which are crucial enabling technologies for software defined vehicles.

Just like with smartphones, users will be able to access a vehicle app to select and download features as their needs change over time, even if they didn't purchase them initially. This could include app enhancements, service enhancements, or functionality enhancements, depending on the user's requirements. This dynamic nature of infotainment systems will allow vehicles to adapt to changing regulations, powertrain requirements, and user preferences, providing a more personalized and convenient driving experience.

Yeah. There's obviously a lot of potential for doing some really neat things. We have one car, and it's an older car and it doesn't have a lot of smart features. But the way that I drive is not the same as the way that my husband drives, you know, as far as accelerating and decelerating.

And that's the cool thing about personalization, right? It can be driver dependent. And I think it's going to come to a point to the very thing you're saying, right? You have a shared vehicle, but you'll be able to set the parameters based on who's driving in the vehicle.

This is where I wonder about tradeoffs in automation and human control. I might have to, or want to, manually go into the infotainment system and say, “Hey, it's me”, versus it recognizing, say, from the weight of the person in the driver's seat and concluding, “oh, this is Karen driving, this is not her husband driving”. Our scale at home does this — my app tries to figure out which profile a new weight reading probably belongs to — but it lets me override that assumption before the data is saved.
Machines can judge a lot from sensors and make inferences, but they’re not always going to be accurate. It's nice to have it automated, but it's also nice to make sure it's right and to have control over it, you know?
There’s the privacy aspect. Should the car company be able to sell the data about what they think is my weight, or interior pictures from the car, to a data broker, or to an AI company for training a model, without my consent?
And what do you do when you get a new driver? It’s a “cold start” for the AI system - it doesn’t know anything about them. For instance, you've got a teenager that just reached driving age, and they're trying to learn. So it'd be nice to maybe configure larger braking margins for new drivers, you know?

Exactly. The concept of multiple seat memory settings, currently available in many vehicles, can be expanded to encompass the entire vehicle, creating a more personalized and convenient experience for each user. Just as users can currently save seat preferences for different drivers, such as themselves, their spouse, or other family members, this same level of customization could extend to various aspects of the vehicle.

With this expanded personalization, each user could have their own saved preferences for things like infotainment settings, climate control, and even safety features. When a user enters the vehicle, they could simply select their profile, and the vehicle would automatically adjust to their pre-defined settings, providing a seamless and tailored driving experience.

Okay. That's a neat vision.
I’m wondering too whether your personal data or content has been used by an AI-based system. You fly a lot, so you're probably going through the TSA and their biometrics and photo screenings. They're using machine learning now to compare your ID to your face, things like that.
So there are probably a lot of ways that machine learning is having an effect on your life. I'm curious if you have any thoughts about any of those travel security features. Or about LinkedIn, and how they're using our content that we put into LinkedIn?

Yes. But I have to admit, one of the things that I've been campaigning for is getting an enterprise LinkedIn account globally. Because LinkedIn is doing some really cool things with AI level technology and really enhancing the experience. Especially on the Sales Navigator side where you can pull in more, I would say, IQ on the account and lead generation. And it can really help you to decipher, what makes your potential lead tick? Where does their interest lie?

And so for me, when you sign up for LinkedIn, with all the data privacy, people don't always realize that their information might be used through LinkedIn. But I think there's a lot to be said about how we can facilitate business, right?

And instead of being afraid of it, I think we just need to figure out, how do we create more transparency and more clarity between the consumer of the data and the one who's contributing the data? So that there's that clear line of understanding of what's being pulled and used in different ways in places. So that way, there's not that mistrust, like, “I don't know what you're going to do with this”.

And for me, in sales and business development, the things I see LinkedIn doing, the lead finder and getting more details around that lead, and what may be a trigger for them, is super important for me in sales, because I'm always trying to close business. If LinkedIn can share a couple of tidbits with me to help me get to a sale, I want to know what those tidbits are. Because I might not readily have the information, and that's the whole purpose of the platform, right? It's really a B-to-B platform, and it's to help me engage with an audience that I may not have otherwise had access to. And so they have unique insights because these people have been interacting on this platform. And that they can share with me to better position myself with that lead or that account. And so for me, I think things like this are invaluable, actually. And I think this is how we really reimagine business going forward into the future.

I don't have any experience with LinkedIn Sales Navigator. I used recruiting features while I was Director of AI at Wind River, because we were actively hiring. So I know LinkedIn Recruiter has some features there. Some were useful, and some weren’t.
I've also seen a recent study about how over 50% of the content on LinkedIn was at least partially AI-generated. So there's some concern about the quality of the content disintegrating, you know?
Another thing that comes up with LinkedIn is there's been some discussion about ageism and how they had some very ageist ads. And it took about 4 months of people protesting to get LinkedIn to take them down. But LinkedIn knows so much about us. I mean, they know what year our degrees were, or our employment history going back. Most people have profile pictures. They can see our network and our DMs. They know a lot about us.
And, again, it's one thing if we opt in and we consent and we say, “Yes, you can use this to help me find colleagues from companies I used to work at”. But if that data also is being sold off to a broker or used for other purposes, then that’s concerning.
So it's interesting to see all the different aspects where AI is impacting our lives. I think LinkedIn has over a billion users, so it certainly has an effect on a lot of people.

That is interesting. That's really interesting, Karen.

It's funny. Some people can tell when AI has been used. I can tell in some cases. I'll read something and know, okay, that person did NOT write that. It’s not their style at all. But as far as using a tool like ChatGPT to help write content, I empathize with people who are not native English speakers who are using these tools to help them polish new content based on their original ideas. I totally understand that.
But sometimes AI-generated content just comes out very bland-sounding, and it doesn't feel original. And often it looks like, okay, someone took a post, and they fed it into a genAI tool and said, “write me a reply comment”, and then they paste it in. And that just doesn't add any value. I can usually tell when someone did that. I read it and go, okay, it’s a waste of time. It lacks original thought, and that is disappointing. If it happens too much, I unfollow or even block.

I agree. That's bizarre. Yeah.

I mean, there's certainly a lot of potential where LinkedIn could do some good things with it. I think that would make it more valuable for us, and it's interesting that you have these insights about the Sales Navigator because I've never seen that. So, good to know that there's some potential opportunities there, especially from a business perspective.

Sales Navigator offers a lot of benefits, particularly in discovering mutual connections between yourself and a potential contact. This tool allows you to identify individuals within your network who are connected to the person you're trying to reach, providing an opportunity to request an introduction through a shared acquaintance. Leveraging existing relationships to establish new ones can be an effective strategy in business development and sales, as it helps expand your network and explore more possibilities. In a competitive business environment, it is essential to exhaust all possible avenues and not overlook any potential opportunities.

And you also obviously want to help your customer solve the real problem that they care about the most.

Yes. Yes.

And that's why those insights are so crucial, right? Like, what challenges are they trying to solve? What pain do they have? What have they been communicating?

When dealing with strategic accounts, particularly publicly-traded companies, it can be beneficial to have tools that can quickly analyze financial documents such as annual reports and 10K filings. This analysis can help identify the company's strategic vision and objectives, as well as understand the information they are sharing with their stakeholders. By gaining a deeper insight into the company's goals and priorities, you can better position your products and services to meet their needs and address their concerns. Accessing and utilizing such tools can significantly improve your ability to target potential customers.

Yeah, the summarization capabilities of AI, I think, are one of the more useful tools. The only danger is that sometimes they end up hallucinating or making up things that actually weren't in the documents. That's a risk. I think the tools have gotten better, and there are ways now to tell it to only use text from primary sources. But you wouldn't want to rely on it and have it actually not be right.

Yes.

I really appreciate you sharing these perspectives, because I don't come from that same business orientation that you do, and you've got the technical background as well. So your insights are really valuable. Is there anything else that you would like to share with our audience today for this interview?

Not at this moment!

Well, thank you for the time - I enjoyed our conversation!

Thank you. Thank you for this opportunity. That's awesome.

Interview References and Links

www.360ofme.com (company page on LinkedIn)

Leave a comment

About this interview series and newsletter

This post is the last 2024 entry in our interview series on “AI, Software, and Wetware”. It showcases how real people around the world are using their wetware (brains and human intelligence) with AI-based software tools or being affected by AI.

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I don’t use AI”:

We want to hear from a diverse pool of people worldwide in a variety of roles. If you’re interested in being a featured interview guest (anonymous or with credit), please get in touch!

6 'P's in AI Pods is a 100% reader-supported publication. All new posts are FREE to read (and listen to). To automatically receive new 6P posts in 2025 and support our work, consider becoming a subscriber (it’s free)! (And if you have professional development funds to spend before year end, take advantage of our 2024 discounts to lock in an annual paid membership.)

Subscribe now

Enjoyed this interview? Great! Voluntary donations via paid subscriptions are cool; one-time tips are deeply appreciated; and shares, hearts, comments, and restacks are awesome 😊

Share 6 'P's in AI Pods

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Thanks for reading 6 'P's in AI Pods! This post is public, so feel free to share it.

📜 AISW #030: Lakshmi Veeramani, India-based AI software leader (AI, Software, & Wetware interview)

Lakshmi Veeramani — Thu, 14 Nov 2024 11:06:54 GMT

Introduction - Lakshmi Veeramani

This post is part of our 6P interview series on “AI, Software, and Wetware”. Our guests share their experiences with using AI, and how they feel about AI using their data and content.

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Subscribe now

Interview - Lakshmi Veeramani

I’m delighted to welcome from India as our next guest for “AI, Software, and Wetware”. Lakshmi, thank you so much for joining me today! Please tell us about yourself, who you are, and what you do.

I’m Lakshmi Veeramani, and I bring 23+ years of experience in the software industry, with a decade dedicated to the field of AI.

Currently, I’m a Senior Architect at Persistent Systems, where I focus on developing the AI/ML/GenAI product vision and roadmap for large-scale projects. My responsibilities include designing the architecture for GenAI agents development and implementing them with AWS cloud-native services as AI/ML/GenAI Engineer. I also have the honor of leading our AI/ML/GenAI squad of my project, where we strive to implement engineering standards and best practices that drive innovation and excellence in our work.

I’m always excited to have discussions with you, Karen. Whether it was our daily one-on-ones during our time together at Wind River or our occasional catch-ups for our blogs, I truly value our conversations.

I do as well, Lakshmi; I’m so happy that you’ve agreed to do this interview with me!
Please tell our audience about your level of experience with AI, ML, and analytics - how you’ve used it professionally or personally, and studied the technology.

I have been professionally involved in AI, ML, and analytics for the past 10 years. However, my journey with rule-based fuzzy logic and neural networks dates back even further, to my M.E. project thesis on lateral autopilot design for aircraft in 1999-2000.

In the early stages of my AI career, I utilized machine learning algorithms for analytics, predictive maintenance, and anomaly detection. I always strive to stay updated with the latest advancements in the AI technology space. More recently, I've been focusing on developing generative AI-based agents.

My work primarily revolves around empowering engineers and professionals to make informed decisions, and I take great pride in that effort.

That’s a great background! Can you share a specific story on how you have used AI or ML? What are your thoughts on how the AI features [of those tools] worked for you, or didn’t? What went well and what didn’t go so well?

Good:

I've found that using ChatGPT for searching has become far more convenient than traditional Google searches. It helps me write code faster and more efficiently compared to searching through Stack Overflow.

YouTube's music recommendations are another feature I love. When I'm in the mood for a particular genre or vibe, I just play one song, and the algorithm takes care of curating a perfect playlist, especially since I don’t always remember all the songs I like 😀.

Google Maps has completely transformed the way we travel. I can distinctly recall life before and after its introduction. It has made long trips safer and more convenient, and I’m excited to see how it continues to improve, especially as more safety features are innovated and integrated.

On my iPhone, the Photos app's AI feature recognizes people. So if I want to create a birthday video for someone, I can simply click on their face, and the app gathers all the photos of them in seconds.

Microsoft Outlook's focus hour suggestions based on my calendar are allowing me to better manage my time. Similarly, Zoom’s automatic meeting notes had been a helpful feature.

The face unlock feature on my phone, and in banking apps like PhonePe, has made access so much easier and safer. Even at my office, checking into the building with facial recognition is seamless.

Those are all great examples of AI in everyday life. And I’m so glad to hear of your positive experience with face recognition. It’s one of the AI application areas that has been problematic for bias and not recognizing people properly.
So those are some things that have worked. Is there anything that hasn’t worked well?

Not so Good

The voice recognition on all the apps search feature has had an unintended consequence. My daughter started using voice commands from the age of three, and now she hardly practices typing, which I feel might hinder her development of spelling and memorizing skills at this stage.

I tried starting a YouTube channel and used various AI tools for content creation, but I found them less helpful than I expected. With so many tools available, it sometimes feels like those who master the tools become successful faster than those with original content but less familiarity with the tech.

That’s an interesting observation about your daughter and using voice commands - that relying on it might hurt her later, due to less emphasis on spelling and typing. Some educators I’ve talked to say that they are trying to find ways to teach their students how to use AI, but without it impairing their ability to learn the underlying skills. It’s a big challenge.
If you have avoided using AI-based tools for some things (or anything), can you share an example of when, and why you chose not to use AI?

I’m really not a fan of AI-powered health apps, especially when it comes to meditation and mindfulness. It almost feels like it defeats the whole purpose. Meditation is something spiritual, something I do for my soul, so relying on an app to guide that experience just doesn’t sit right with me.

As a parent, one of my biggest challenges is managing my daughter’s screen time. Ideally, I’d love to avoid giving her the iPad altogether, but it’s tough—especially in a world where devices seem to be everywhere.

I’ve read so many great books throughout my life, and I cherish those experiences. But as a mother, I sometimes wonder, will my daughter ever have that same relationship with reading? These days, she’s glued to her iPad from morning to night, and it feels like AI has learned how to capture the attention of kids and teens so effectively. It’s a different challenge than what our parents faced.

Part of me just wants to let her grow in her own way, without pushing her too hard into the “work hard” culture I grew up with. The truth is, I don’t even know what the future job market will look like in 10 years. AI is reshaping the workforce at such a rapid pace, I find myself questioning what kind of dreams I should encourage her to pursue. What does success even look like in a future that’s so uncertain?

And I feel like this is a challenge that parents who work in AI face more than others. We’re keenly aware of how this technology is evolving and how it might impact the next generation. It’s a lot to think about, and honestly, it’s something I’m still figuring out as I go.

I think parents everywhere are figuring that out as they go! 🙂
I know what you mean about reading. I’ve been an avid reader since I was 5. I do like being able to have a new book with me ‘anywhere’ on my phone, nowadays, but I still buy some printed books. It’s a different experience. I remember when e-books and e-readers first became popular, and people were worried that printed books would go away. They haven’t.
A common and growing concern nowadays is where AI and ML systems get the data and content they train on. Some have been trained on copyrighted books that were simply stolen. Some have used data that users put into online systems or publish online. And companies are not always transparent about how they intend to use our data when we sign up.
How do you feel about companies using data and content for training their AI/ML systems and tools? Should ethical AI tool companies be required to get consent from (and compensate) people whose data they want to use for training?

One of the things that’s been on my mind is how AI, especially large language models, are trained on a vast amount of data without necessarily asking for anyone’s permission. This is especially troubling for the art and creative communities, who I think will face some challenges in the short term. But over time, I believe original music will prevail. I can’t imagine people getting excited about a song that was announced as “composed by AI”.

From my own experience growing up watching Tamil movies and listening to music, it’s often hard to predict which song will become a hit. There’s no one-size-fits-all formula for success. Sometimes, fan loyalty to actors even drives the success of music, whether or not the songs themselves are that great. So, at least in my view, Tamil fans—and many others with deep cultural roots in their music—won’t easily fall for AI-generated music.

In the words of the famous Tamil poet Avvaiyar, who once said “Padal pudhiyadhu”, meaning “Songs are new”, I interpret this as music needing to feel fresh. If it’s just learning from the past, where’s the newness? (We could dive deeper into that, but I’ll leave it for another time!)

All great points - as you know, I’m passionate about ethical use of AI for music 🙂 And I know there have been a handful of AI-generated songs which have built up a lot of ‘streams’. It’s hard to know how many of those were just curiosity about AI vs. people thinking, “Hey, this song is really good, I need to tell my friends”. Or how many would come back to listen to another song created by the same prompter.
The other thing I’m seeing is that the streaming platforms are being overrun by AI-generated music. Most of it’s considered to be mediocre. And the AI platforms that scrape are going to be training on that ‘new’ mediocre input. It seems like it could be a race to the bottom for music quality. As you said - let’s save that topic for another time :)
There are so many areas where AI ethics are critical, and music is one great example. I’m sure you have more 🙂

Yes, Karen, I really appreciate your passion towards the ethical use of AI in music and the work you do!

As a working woman, I also have concerns about AI-driven talent acquisition tools. What if these systems carry biases, especially against women? We’ve fought long and hard to get where we are, and the last thing we need is to face discrimination from technology itself. It’s a real worry.

That’s absolutely a fair concern. In fact, you probably saw the recent report on Workday’s AI hiring tool being biased. (link)
AI is sometimes called a ‘stochastic parrot’ and it easily parrots the biases of the data it was trained on, and the people who trained it. Even when people are conscious of bias and try to mitigate it, it’s challenging to do. So yeah, AI bias in hiring and HR tools is potentially very serious.

In the software world, I’ve noticed that while we used to have just a few star coders in a team, now, with AI, almost everyone can code competently. The difference is going to be how creatively you approach problems will make you star in the current scenario. It’s not about learning new programming languages; it’s about thinking innovatively. And that’s where we, as humans, will continue to outshine AI. After all, AI is learning from what humans have already done—it’s not truly creating something new on its own.

Exactly - data quality and bias checking and ethical sourcing are paramount. And innovative thinking and problem solving are always going to be essential. Getting the problem to be solved identified and well-defined is critical, and we haven’t yet seen AI able to handle that.

A related concern that I don’t think we’ve fully addressed is how we as people leaders develop our software teams, how we grow beginning developers into experienced senior developers and leaders. Most junior people learn from struggling with progressively bigger, harder problems and being mentored by senior people. Using AI for ‘coding’ seems to disrupt that growth pattern. People ‘write’ the code, but they don’t necessarily understand the system they’re building.

That’s a great point - we need to figure out how new developers using AI-based tools can build the expertise to understand AI-generated code and systems well enough to know when they’re not designed or coded right.
Transparency seems to be a universal demand from people who want to have that kind of visibility and control over how their personal data is used.

Definitely, Karen - there's this saying we hear a lot these days: "Data is the new oil." And I think it's true—anyone who’s making a commercial product using the data should absolutely pay for the data they're using.

When you’ve worked with building an AI-based tool or system, what can you share about where the data came from and how it was obtained?

Personally, I wouldn’t create any tools based on public domain data. The work I’ve done so far is industry-specific—things like manufacturing, engine data, battery data—and none of that involves human data.

That makes sense. One of my first guests, Ralf Gitzel, described something similar, about how he and his team had to invest a great deal of effort to collect the machine data they needed for their models.

Good to know, I will go through that, I believe that when AI is used wisely, it can really help people. A lot of industry-specific tools are reducing human effort, improving safety, and enhancing lives in many ways.

I agree!

When we’re developing AI solutions, one of the key things to keep in mind is following GDPR compliance. Now, while GDPR technically applies only to the EU, it’s actually a really good framework to start with for any region, as I mentioned in my article about The Lost Soldiers and the Map, as some map is better than no map, when you are lost.

But beyond the penalties, designing AI apps that are GDPR-compliant puts us on the right path toward building “privacy by design” AI systems. This approach not only helps us meet current regulations, but also prepares us for any new rules that might come up in the future. More importantly, it ensures that the AI we’re building is ethical, respecting user privacy at every step.

Absolutely - that’s so important. Yet, as members of the public, there have been cases where our personal or private data or content has been used by an unethical AI-based tool or system. Do you know of any cases that you could share (without disclosing sensitive personal information, of course)?

Have you ever noticed how you browse something online, and then suddenly, you’re seeing ads for it everywhere, even in other apps? It happens to me all the time.

Yes! Did you hear Tracy’s and Quentin’s stories about this? It seems to be very common nowadays - not only what you browse online, but even things you do and say ‘offline’.

Yes, I have listened to Quentin’s stories. And it doesn’t stop there—back when I used to walk every day from 5 to 6 PM, I’d get all these calls for weight loss programs, right around that time. I blocked those callers, and thankfully, I’ve stopped getting the calls, but it really made me wonder. Is my personal data being shared with these advertising agencies? Whether we like it or not, it feels like AI-powered ads are always following us, and to be honest, I really don’t like that.

I don’t blame you! It’s creepy. I looked into this a while ago (link) and it seems that a lot of phone apps are grabbing our data. It’s usually buried in those tiny terms and conditions. Even having push notifications enabled can give an app our location data, because our location is in the metadata for app notifications. So it’s really hard to avoid and protect yourself.

In India, it’s even worse. They ask for your phone number for every little transaction, and the next thing you know, your number is being passed around to advertisers. I must’ve blocked at least hundreds of calls by now from the same firm. Even today, if I give my number at a store for billing, it’s pretty much guaranteed I’ll start getting spam calls. It’s like we’ve all just gotten used to it.

I’ve had a similar experience in US grocery stories, where the clerks just asked for your phone number like it’s normal and expected, and implying that it isn’t optional. I finally just started saying “I don’t give it out”. At first I got weird looks, but they kind of shrugged and said “ok”. And I’ve been hearing about more people doing this.
We (especially us women) have to learn to not be so nice and agreeable, and push back when people and companies ask us for information that they don’t need. Because as you said, they’ll use it and sell it, and it’s generally not for our benefit; it’s for theirs.
The other spam calls I get nowadays have mostly been political. I was getting 10 or more a day coming up to the November election. My mobile number is not in my voter records, so they probably bought my info from some other source and matched it up. Or maybe they are just robo-dialing all of the possible number combinations in an area code. 😉 Either way, they’re never going to reach me, because I ignore all unknown callers, and I block all junk texts about politics, no matter what side they’re supporting. And my phone now has gotten pretty good about flagging them as ‘possible spam’ - so a ML model somewhere is likely taking care of that.
Do you know of any company you gave your data or content to that made you aware that they might use your info for training AI/ML? Or have you ever been surprised by finding out that a company was using your info for AI? It’s often buried in the license terms and conditions (T&Cs), and sometimes those are changed after the fact. If so, did you feel like you had a real choice about opting out, or about declining the changed T&Cs?

That’s good to know what you did! What I really think is that, by default, apps shouldn’t be collecting our data. Right now, most apps collect and use our data unless we actively opt out—and that’s just wrong. It should be the other way around: apps should need our explicit permission to use any of our personal information.

Absolutely. And I envy the people in the EU who have GDPR to protect them from this. The rest of us don’t - yet.

Yes, it is good that the EU has GDPR. And then just last week, Facebook got into trouble for storing passwords in plain text in Ireland. They got sued for it.1 And it’s shocking that these big companies are still making such basic mistakes when it comes to data security.

Yes - that’s awful! It’s not really surprising any more that public distrust of AI and tech companies has been growing. What do you think is THE most important thing that AI companies need to do to earn and keep your trust? Do you have specific ideas on how they can do that?

Trust is something technology companies have really worked hard to build in other areas. Take Amazon, for example. We’re so comfortable buying online, paying the full amount upfront without a second thought about whether we’ll actually receive our order. That trust didn’t happen overnight—it was built because they’ve consistently compensated customers without making it an exhausting process. The same goes for companies like Zomato, Swiggy, or BigBasket. If there’s an issue with the product, they’ll refund or replace it, no questions asked. That kind of customer-first approach has built a lot of trust.

You’ve put your finger on exactly one of the key issues, I think. The companies that exploit our data are financially motivated to manipulate us into not making more conscious, more frugal decisions.

A recent paper I came across from the Harvard Business Review, AI’s Trust Problem, really hits the mark on this.2 It talks about how there’s a trust gap between humans and AI that needs to be addressed. The paper outlines 12 key AI risks that are frequently cited, and I totally resonate with it. These risks include:

Disinformation
Safety and security
The black box problem
Ethical concerns
Bias
Instability
Hallucinations in large language models
Unknown unknowns
Job loss and social inequalities
Environmental impact
Industry concentration
State overreach

These are serious challenges that AI product companies need to focus on if they want to earn our trust in the same way companies like Amazon have.

Great information! (Readers: a link to the HBR article is in the end notes).
Lakshmi, is there anything else you’d like to share with our audience?

As we navigate the delicate balance between life and death, what truly keeps us alive is the profound sense of growth. Humanity's innate drive to innovate—whether through AI or other technologies—will never cease. However, every invention must be a force for good, uplifting humanity and ensuring we all move forward together.

AI is doing lots of good as well. So I would like to keep the attitude of doing good with the AI power, and all you need is the AI Wisdom!!

Leave a comment

Interview References and Links

Lakshmi Veeramani on Medium (see her article on Responsible AI)

Lakshmi Veeramani on LinkedIn

Lakshmi Veeramani on Substack

About this interview series and newsletter

This post is part of our 2024 interview series on “AI, Software, and Wetware”. It showcases how real people around the world are using their wetware (brains and human intelligence) with AI-based software tools or being affected by AI.

And we’re all being affected by AI nowadays in our daily lives, perhaps more than we realize. For some examples, see post “But I don’t use AI”:

We want to hear from a diverse pool of people worldwide in a variety of roles. If you’re interested in being a featured interview guest (anonymous or with credit), please get in touch!

6 'P's in AI Pods is a 100% reader-supported publication. All new posts are FREE to read (and listen to). To automatically receive new 6P posts and support our work, consider becoming a subscriber (it’s free)! (Want to subscribe to only the People section for these interviews? Here’s how to manage sections.)

Enjoyed this interview? Great! Voluntary donations via paid subscriptions are cool; one-time tips are deeply appreciated; and shares, hearts, comments, and restacks are awesome 😊

Share 6 'P's in AI Pods

Series Credits and References

End Notes

Meta was sued and fined for storing passwords in plaintext and allowing them to leak. The password exposures happened in 2019 and DPC Ireland fined Meta for it on Sept. 27 of this year. https://cybernews.com/security/meta-100m-fine-dpc-ireland-plaintext-passwords-facebook-leak/

“AI’s trust problem: Twelve persistent risks of AI that are driving skepticism”, by Bhaskar Chakravorti / Harvard Business Review, 2024-05-03.

6 'P's in AI Pods (AI6P): People & Places

📜 AISW #105: Joan Kinyua, Kenya-based digital rights advocate

Introduction - Joan Kinyua

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Interview - Joan Kinyua

Interview References and Links

Data Labelers Association website and on LinkedIn

About this interview series and newsletter

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

📜 AISW #100: Karo Zieminski, Denmark-based product leader and builder

Introduction -

Note: In this article series, “AI” means artificial intelligence and spans classical statistical methods, data analytics, machine learning, generative AI, and other non-generative AI. See this Glossary and “AI Fundamentals #01: What is Artificial Intelligence?” for reference.

Interview -

Interview References and Links

Karo Zieminski on Substack (Product With Attitude)

About this interview series and newsletter

Series Credits and References

Microphone photo by Michal Czyz on Unsplash (contact Michal Czyz on LinkedIn)

Credit to CIPRI (Cultural Intellectual Property Rights Initiative®) for their “3Cs' Rule: Consent. Credit. Compensation©.”

Credit to for the “Created With Human Intelligence” badge we use to reflect our commitment that content in these interviews will be human-created:

📜 AISW #093: Elettra Fiumi, Swiss-based Italian-American AI cinema innovator

Introduction -

Interview - Elettra Fiumi

Interview References and Links

Fiumi Studios (fiumistudios.ai)

Elettra Fiumi on Substack (AI Cinema With Elettra Fiumi)

About this interview series and newsletter

Series Credits and References

📜 AISW #087: Jean Gan, Singapore-based Senior Legal Manager

Introduction - Jean Gan

Interview - Jean Gan

Interview References and Links

Global Legal AI and the Global AI & Law (GAIL) Network

About this interview series and newsletter

Series Credits and References

📜 AISW #084: Emily, USA-based computational memory researcher

Introduction -

Interview -

Interview References and Links

Emily on GitHub (direct contact)

Emily on Substack (The Nth Dimension)

About this interview series and newsletter

Series Credits and References

📜 AISW #082: Mia Kiraki, Romania-based founder & CMO at Yahini and content creator & strategist

Introduction -

Interview -

Interview References and Links

Mia Kiraki on Substack (Robots Ate My Homework)

About this interview series and newsletter

Series Credits and References

📜 AISW #078: Jeremiah Cioffi, JD, MA, USA-based attorney and AI entrepreneur

Introduction - Jeremiah Cioffi, JD, MA

Interview - Jeremiah Cioffi

Interview References and Links

on Substack

About this interview series and newsletter

Series Credits and References

📜 AISW #077: Sam Illingworth, Scotland-based university professor and writer

Introduction -

Interview -

Interview References and Links

on Substack : Slow AI

About this interview series and newsletter

Series Credits and References

📜 AISW #073: Dee McCrorey, USA-based transformation strategist

Introduction -

Interview -

Interview References and Links

on Substack ()

About this interview series and newsletter

Series Credits and References

📜 One Year of AI, Software, and Wetware: What's Changed?

A Year?!

Your turn - poll open to all subscribers!

About this interview series and newsletter

Article References and Links

Kathy Gerstorff on Substack