Introduction - Kadri Adekunle Mayowa
This edition of AI, Software, and Wetware features an audio interview with Kadri Adekunle Mayowa, a 🇳🇬 Nigeria-based product generalist. We discuss:
His professional experience applying AI and machine learning to solve business problems in the finance industry
Why he doesn’t rely on generative AI tools for communications
Pros and cons of ML models for predicting customer churn
How laws in Nigeria protect consumers rights to privacy and data protection
The “Check AI” Chrome extension he built for his AI safety project to detect biases in chatbot answers
and more. Check it out, and let us know what you think!
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 interview is available as an audio recording (embedded here in the post, and later in our AI6P external podcasts). This post includes the full, human-edited transcript. (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.
Interview - Kadri Adekunle Mayowa
Karen: I am delighted to welcome Kadri Adekunle from Nigeria as my guest today on “AI, Software, and Wetware”. Kadri, thank you so much for joining me on this interview! Please tell us about yourself, who you are, and what you do.
Kadri: Alright. Thank you for having me, Karen. My name is Kadri Adekunle, as you mentioned. I’m a product generalist based in Nigeria. When I say generalist, I mean I’ve done a lot of stuff: building software products, from driving good for this product, driving analytics for this product, also on the technical side of things. I have 10 years of experience in telecoms, financial technology, startups, and e-commerce.
I see it at the intersection of products, growth strategy, and analytics. Just look at that guy that wants to turn complex problems into scalable solutions. I’ve worked with a lot of teams, worked with about 150 people, launched products around $1 million, $7 million in recurring revenue.
Beyond this, I’m always passionate about mentoring people, community leadership, and sharing knowledge. I’ve spoken in about five tech events. I’ve just been supporting different products and data leaders in Africa.
Karen: That sounds great. Thank you. It sounds like you’re mostly in what a lot of people would recognize as product management. Is that fair? But you also develop software yourself?
Kadri: Yeah. Yes.
Karen: Okay. You do it all!
Kadri: Yes. Generalist, as I said.
Karen: Good. All right. So tell us a little bit about your level of experience with AI and machine learning and analytics. I’d like to hear how you’ve used it either professionally or personally, or if you ever studied the technology.
Kadri: All right. I have professional experience applying AI and machine learning to solve business problems. AI has always been there. It may not be generative, but then some of us have been predicting next outcomes using artificial intelligence. It was not popular. It was not mainstream. It was not like in the chat era, you send AI agents work and they do it for you. But it has always been there. And some of the experience was when I was working in AXA Mansard. AXA Mansard is an insurance group in the world, and they also sit in Nigeria. I remember having issues. When I say issues, I mean, when people want to make claims, there’s always this backlog. Because a lot of people drive rough in Lagos. So one minute you buy your car, the next month someone has hit you on the bridge. So you see a lot of requests. But then we don’t have enough manpower to support them.
So we thought about this. We put some requests in a bond. I mean the amount that it’ll take you to fix your car. So in that bond with between zero and hundred thousand, we wanted it to be a straight-through process. So what happens is there’s an AI-powered damage detection solution that will analyze your claims and put them in that amount corresponding with the amounts that you have uploaded. It’s like a computer vision and machine learning solution that analyzes vehicle images. So it assesses the damage and it’ll generate the repair estimate, then start applying our claims process. All of those manual processes were reduced.
Also, most recently working in Access Bank. It was seated in London, seated in more than 16 countries. We built an AI predictive model to anticipate churn. That’s what I’m working on right now. And I also built one within our contact center team to forecast manpower and optimize our staffing levels. So I have done stuff with AI in the past and currently using it, and it helps.
Karen: Yeah, it’s really interesting what you’re describing with the use of computer vision for the claims handling. I think a lot of people, when they talk about AI, they think just about generative AI and what’s come out since late 2022. And AI itself has been around for about 80 years in some form or another. Most of my experience also was prior to generative learning coming out, and working with things like image recognition. I’m sure we’ll have some good conversations about that. So thank you for sharing that.
How did you learn about machine learning? Did you study that in school or did you learn it on your own? What worked for you to learn the technologies?
Kadri: I had a background in engineering. Considering our school then, we were very big on the theory, not on the practical side of things. But then when I finished school, I wanted to fit the work model. So I thought to go and learn different programming languages, right? I started sharpening my skills in Excel. Then at a point, I learned how to use SQL. Then I started with Python. So I have a moderate knowledge on how to use Python, even though I have a team that does all the dirty jobs for me now. But it started when I learned how to use Python to wrangle data and just make sense of data. So my journey started from there.
Karen: Good to hear that. It sounds like you’ve used AI quite a lot professionally. Do you ever use AI tools in your life outside of work?
Kadri: Yes. I use AI tools outside of work. There are times where I want to support my wife, who is a business owner, right? So she gets a lot of DMs. All of these small, medium enterprises that use Instagram as their mode of communications, trying to convert. They do advertisements online. So you’re chatting, you’re missing a lot of chats. A lot of people are trying to talk to you at the same time. A lot of people are interested in your items.
So I thought about it that we could automate things. We have all of these API access from Instagram addressed, so we’re able to automate some of those. They’re simple things anyways. So AI would help you to figure out the response, so whatever comments they make in direct messaging, was able to reduce turnaround time on requests. We were able to automate a lot of things that needed to be automated. So that’s one.
Two is me trying to send an email, maybe not even work-related. Maybe I need to tweet online, then I need to do some research. So outside of work, here has really been helping. But then I’m just one of those that would not. Because sometimes it makes you lazy in the brain. So I try to find a balance between always going to AI for help.
Karen: That’s a really good observation. It’s like using a GPS to navigate and get around your city. They’ve seen that people that rely on it more lose their ability to form a mental map of the region and remember streets and locations and such. But it’s extremely convenient if you’re in an area that you don’t know. It’s somewhat similar, I think, with people using bots to answer questions.
Kadri: Yes. Thank you.
Karen: I’d like to hear if you have a specific story on how you’ve used one of these tools. It can be AI or machine learning features. I’d like to hear about your thoughts on how the AI features of those tools worked for you or didn’t. Basically, what went well and what didn’t go so well with it?
Kadri: All right. Thank you for that question. I like that. So I currently work in financial services, right? One thing on our partner right now is predicting churn. So in the past we just categorized inactivity into buckets. Okay, this one is actually three transactions. This one has just done maybe two transactions in three months. This one has not done any in six months. And we dig deeper into their behaviors. But right now what we’re using machine learning to do is predict people that are about to turn, maybe, in the next 30 days, 60 days, 90 days, right?
So it helped us. We use a lot of features, like their transaction behavior, product usage complaints, just engagement signals to put them into a bucket, using a probability score, right? So I think that’s something that AI’s feature has done for us. So we anticipate churn. Obviously, technology will always fail you. People will always complain. People will compare you with competition, no matter how you try. For us, we anticipate it. Doesn’t mean it’s a bad thing. Doesn’t mean that we’re not doing our jobs well. But sometimes it’s what it is at the end of the day. In that regard, AI also helped.
The model itself helped us. It’s like an early warning signal, right? Something that was not obvious during traditional reporting, like a drop in transaction frequency. So it helps us just come up with a decision to say our churn recovery strategy. Where do we improve? What do we act on? What are the next steps? Just more of resource prioritization. Where are the quick wins? How do we get better? Those are the key areas that I can remember now in terms of our AI features and how it has helped us.
And what has not gone well is also around maybe data quality sometimes. The key drivers of churn, it’s usually harm complaints. Sometimes you don’t get to see some offline reactions, right? Or price transactions. So just the accuracy, something that doesn’t really go well. And sometimes some people don’t want to trust your data, depending on who worked at it. I think that answers your question.
Karen: Yeah. One thing that tends to come up with machine learning models is whether they are explainable or interpretable. Did you have any experiences with interpretability or trust or explainability of the machine learning models that you created for analyzing this data on churn?
Kadri: Okay. On the contact center prediction forecast, the goal of that project using machine learning was because we were having high abandon rates on our queues. When people try to engage us, in terms of email, you see that there are delays, right? Maybe there’s a general downtime, so you see a lot of customers trying to reach you at the same time. So it can increase your abandon rates, your speed to answer, and all of these KPI that have been tracked at the contact center level. So what we did was just to build a prediction to say, “Okay, based on historical data, build a contact center forecast that can predict manpower, the volume of interaction.”
The first attempt at these predictions were all right. But those activities happened very recently, so we try to compare to look at the prediction, but they were far, right? So think about this. Let’s say I beat it two November last year. Then I’m like, okay, fine. Let me see what happened. So October has happened, people have tried to reach us, so what are the volumes that we get? So when we use that AI, those predictions were not accurate. So we had to retrain the data. We had to correct a lot of nuances, a lot of outliers. But eventually the model was correct and all of our stakeholders trusted us.
Karen: So you’ve made good use of AI and machine learning tools. I’m wondering if there are some things, or any thing, where you would avoid using AI or machine learning, and why you would choose not to use AI for those purposes?
Kadri: Alright, so for me, especially in the official setup, I hardly use AI for communicating, because there’s this tone when you send messages via AI, there’s just this disconnect. There are some times that once you see it, there’s really nothing you want to tell me about these columns. I bet it was written by an AI, which is fine, which is not a problem.
But personally, in my own opinion, I just don’t want to be weak mentally. In terms of communicating, I don’t want to be lazy. I want you to feel my intent. Sometimes there will be typos. Maybe we want to type very fast. I would prefer you see my typo, to be very honest. I don’t know if I’m normal, but that’s just how I think. Because you see empathy in the communications. You see original thinking, right? It’s not the easy way out. So that’s the way I see it. It makes it more authentic, right? But then it is what it is. It makes our jobs faster, but on a personal view, I don’t rely on it completely to communicate.
Karen: A lot of the AI-generated writing, it’s harder to tell now sometimes. But there is still a sense of the human behind it. I think most of us want that authenticity to come through regardless. So it’s interesting that you avoid it for that kind of writing. And obviously anything that you would write, you need to be accountable for, right?
Kadri: Yes.
I think use of it for catching typos is pretty common and people hardly even think of that as AI or machine learning nowadays. I mean, we had spelling checkers long before generative AI came out, so I don’t think anyone minds it being used for those purposes. A lot of people are writing in a language that isn’t the one they grew up speaking in. Having some assistance with that, I think, most people don’t mind. But having it write an answer to this email for me, even if it’s not obvious from the word or the phrasing – it feels hollow, you know?
Kadri: Yeah. I get it.
Karen: One question that always comes up is where these AI and machine learning systems get the data and the content that they train on. A lot of times, especially with generative AI systems, they’re using data that we have put into online systems or some creative works that people have put online. And a lot of companies aren’t very transparent about how they intend to use our data when we sign up for their services. So people can get surprised about how their data’s being used. There’s a concept called the three Cs, which is from CIPRI. It’s the idea that creative people are entitled to consent, credit and compensation if their data or their works are used for training a system or a tool, and I’m wondering what your thoughts are about that. Do you think it’s okay for them to scrape whatever is what they call publicly available, even though it’s not public domain? Or do you feel like the three Cs are important if anyone’s trying to use your data or anything that you’ve created?
Kadri: Yes. I think the 3Cs is something that we should all adhere to. Some companies do that. But then I think companies need data, at the end of the day, to build AI systems. But then the way it’s collected and used today is often not very clear to us. People don’t always understand that their content, all of those interactions, are used to train AI systems. Not all companies are always transparent about this. Some are. But that lack of clarity creates a trust problem. So there’s usually a gap in how these data are collected.
Because of this, I believe ethical AI companies should clearly explain how data will be used and give people real choices, including maybe if they want to opt out, right? So for user-generated content, consent should be the baseline. And there should also be a fair compensation, which should always be considered, especially when the content creates commercial value.
There’s even an AI safety project that I worked on in the past, a project that involves building a Goku Chrome extension. It was designed to detect bias in AI-generated responses and surface those biases in real time. So the goal was not just to flag issues after they come up, but to help users recognize and address biased outputs as they happen. Working on these reinforced learning projects introduced us to training data and how important transparency is in the real world.
At the same time, I don’t think all data should be treated the same. Public, anonymized, and aggregated data can reasonably be handled differently from private or sensitive data. What matters most is fairness and clarity, right? An AI system will be socially looked up if people trust it and how their data is being used. So I think that answers the question.
Karen: Yeah, that was very interesting to hear. I’ve got a couple follow-up questions. You mentioned that the default should be for people to opt in and they shouldn’t have to opt out. I completely agree with that. I’m wondering, what are the laws in Nigeria right now that pertain to people’s rights to be able to opt out?
Kadri: Alright. For example, sometimes we want to do a customer research program within financial services. There are a lot of NGDPR rules, right, that govern how we use data, how we collect data. What do we use your data for? So the first thing for us is we let them know the intention. It’s like a data protection regulation, right? It governs how personal data is stored and used and also shared. So the principles that I remember is that there should be lawful and fair processing. You should also let them know that there’s consent. We always update consent before collecting your data. That’s the first thing you should see.
We also let them know the reason. So it’s like a data privacy law that is embedded on our websites. So what was the specific reason why we’re collecting this feedback or this data, right? We also only collect the data that is necessary. So we protect data against loss, thefts. Those are rules.
And there are data subject rights, too. Individuals can always access their data. They can request for correction, they can withdraw consent. If you don’t want to get some type of communications, you can always opt out.
So there are a lot of laws for banks, schools, hospitals within Nigeria. If, let’s say, that you are using my data for a marketing campaign and I didn’t opt for that, I can take it to the court. Calls that I don’t expect. Text messages. So there are laws guiding data. We may not be there yet in terms of a global view, how data are trained, managed, and people compensated. But then we’re evolving into that as we speak, because there are a lot of communities, a lot of education across banks.
Karen: So Nigeria has an equivalent to the GDPR. In some ways, you’re actually better off than we are here in the US in that regard.
Kadri: Yeah. Yeah.
Karen: I’m also curious about what you mentioned with your AI safety project. You built a Chrome extension to detect bias in answers that were coming from chatbots, is that right?
Kadri: Yes.
Karen: That is so interesting. I would think a lot of people might want to use that. Is it a publicly available Chrome extension?
Kadri: No, it is not publicly available. It was just a project. I think on my GitHub, but it was not publicly available. It was way, way back in early days, I think maybe four years ago, three years. But maybe I can get my updates here and maybe do something after today’s call.
Karen: Yeah, I’ve been following some writers on Substack who are working with different types of analytics and AI and using these tools for writing code. And some of them are talking about writing Chrome extensions because it can work within, you’re already logged in, so they don’t need to build authentication mechanisms and such. One that could actually detect, “Hey, you know what this bot just told you? You asked it to tell you about a specific kind of fact, and it gave you a biased answer.” – I think that would be really useful.
Kadri: Yes, yes. It’s a nice to have – a must have, to be honest.
Karen: Yeah. Because most of the chatbots were trained on data from our biased world, and so therefore they have biases built in. Maybe sometimes we recognize that, sometimes we don’t.
Kadri: Exactly.
Karen: They’re very heavily biased towards English and toward Western frames of thinking and everything else. There’s biases against women, biases against people of color; from the global south are definitely very much underrepresented. I’ve seen a lot of really interesting and kind of funny, but funny in a bad way, examples of bias coming out of those tools.
Kadri: Exactly, exactly. There are a lot. Maybe this is our time now. You are opening my mind to possibilities. I think maybe I can dig deeper. Maybe one day, some AI company in America will just buy my products, and I move on to other things.
Karen: Yeah. So I’m curious, which of the common chatbots that are out there have you tried?
Kadri: Okay. Yes. I use ChatGPT. I think people underrate ChatGPT so much, right? In terms of the “garbage and garbage out”, once you get your prompts right for ChatGPT, I think it helps my work. Not communications, but just me thinking broadly and letting it pick from maybe other people that have done it in the past. Obviously that’s what it does. I get those answers that they’re not ready-made. There’s still a lot of work to be done, but I think out of all of these chat boxes, it’s something I can relate to.
I don’t know if I’m being biased or I’m so used to it, but I just feel that the content these days makes sense a lot. I also use GenSpark. Maybe I need an idea on how to prepare a slide. So you create them. But not that I’m using this slide, so I’m just going to use it like an inspiration, right? So it makes my work faster. I’m not thinking too much, so I’m just building in PowerPoint.
So those are the two. Sometimes I use Claude for the quotes, maybe to validate something. Maybe some quotes are not running. AI will just tell you, “Oh, there’s a comma missing.” Something like that. “You have to put a space” and all of this. So those are three that come to mind that I really use: ChatGPT, Claude, and GenSpark.
Karen: Have you ever tried any of the versions of these tools that generate images or music or videos, or anything else other than words?
Kadri: Yes. So I’ve used ElevenLabs for, I can’t remember, but it was something around voice cloning. So I wanted to use it, but I can’t remember what my issue was then. I think I knew about ElevenLabs, that they do great stuff. Even for GenSpark, outside the chat, just like I said, you can build slides. You can communicate to it by voice. You can build mini agents that can connect to your Gmail and all of those. Maybe you have an email that you need to read - it just gets them for you, right? So those are the use cases for me in terms of ElevenLabs, ChatGPT, and GenSpark.
Then for image also, I was working on an open source project to design a flyer for the community, like a certificate for those that do something. I can’t remember what exactly, but I go to it to create images, right? I don’t use the image, I just use it as an inspiration to maybe design on Canva or something else. So GenSpark really does it for me.
Karen: Okay. I am not that familiar with GenSpark. I’m going to have to look it up after our call.
Kadri: You really need to. So GenSpark, in terms of chats, they’re not there yet. But you’d be surprised that in terms of the other things that they do on that platform, it’s really, really high-level. I think it’s owned by the Chinese, I’m not sure. But it does great stuff. Image processing, developing, writing codes, designing. There’s AI images, there’s AI videos, maybe your meeting notes. It does all of those, right? So it’s really a very great tool, especially for my slides. It works. You need to check it out.
Karen: Yeah. Sounds good. So you’ve mentioned a lot of different tools that you use that have AI, and they’re obviously based on some sort of training. Do you feel like the companies that provide those tools have been transparent about sharing where they got the data that they use to train those tools, and whether the people whose data they used were given the option to consent to it being used or not? Or do you feel like they have not been transparent?
Kadri: Not at all. I think some have been transparent. A lot of these companies don’t really tell you, right? For example, like people in Europe, they’re covered by GDPR. I’m sure they are able to opt out of, maybe, Meta using their Facebook and Instagram content. Like they train ai, but then, can they opt out? I’m not sure they can opt out on some content. Especially for even Nigerian builders, they don’t really tell you about what they use those data for. Sometimes you get targeted emails that you don’t even subscribe for, right? So they’re not really getting consent from people. Yeah.
Karen: You mentioned that your wife is a business owner, using Instagram for her business, I think. Is that right?
Kadri: Yes. Yes. You’re correct.
Karen: Yeah. Yeah. Meta is one of the companies that we’ve heard a lot about them not giving people the rights to opt out or just saying, “Yeah, if you use our product, you pretty much have to agree that we’re going to use your data however we want.” And that doesn’t sit well with everybody.
Kadri: Yes, sure, sure.
Karen: Yeah. All right, so overall, you feel like the companies aren’t disclosing very well. It sounds like in the companies that you’ve worked in, where they’ve been involved with using data, that they’ve made efforts. But for the tools that you’re using, for the most part, they haven’t. For the tools that you’ve worked on, building some of the predictive models – I think I know where the data probably came from, but could you say a little bit about that?
Kadri: Yes. So for us, it’s very straightforward. We have different data points, right, within the organization. There are a lot of processes, protocol that can even frustrate you if you’re not very strong, because it is data at the end of the day, right? Something you can think is worthless, like conversations. Maybe the timestamps, the call handling, the wait time by hour. Maybe when customers try to engage you, like, when I call you. Those call waiting times, maybe, for the people that interact with customers online, social media, and live chats, their speed to answer all of those data. You need to see the process to get them. Even the conversations that happen with the customer and maybe an agent or a staff, right?
I think in Nigeria we’re very, very big – especially where I’ve worked, AXA Mansard and Access Bank – we’re very big on data protection. You don’t just have access. You can’t even have access to transaction data if you are not getting some things. The group head, the final head, is the only one that can approve those assets. So it’s obviously set in, and you can only even access them on the bank network. Outside of the office, you really can’t do anything.
So we are very big on data protection. Whenever we want to collect it, we also let them know about how safe it is, what we want to use it for, and we don’t over-collect. Even for our surveys, we don’t over survey. We take time to analyze the data to be sure you are one person. You’re different. You’re not having solving fatigue. So we’re very big on data protection, especially where we’re involved in Nigeria.
Karen: That’s great to hear. Thank you for sharing that. I’m wondering, as members of the public and consumers, our personal data or the works that we’ve created have probably been used by AI-based tools or systems. Do you know of any cases that you could share where your data may have been used?
Kadri: So for me, I write publications online, right? I just publish them. Sometimes people reach out to me to just write articles about work, about personal stuff, right? So I’m very certain that this is one way people might collect some of my ideas that I put online, maybe on Medium, Substack, or whatever.
So one common example is also the image generation systems. We have our articles online, our social media content online. These data, they’re publicly accessible. Individual creators are often not notified when their content is included in large training data sets. It’s not like I get a notification that somebody has picked some elements of my picture online, to train data. So obviously there are a lot going on at background, which companies need to be clear about.
Another example is the facial recognition systems. One is craft, like images online. Sometimes they use this to be large biometric data sets. In some cases this can lead to regulatory actions and legal challenges due to concerns about consent and privacy. So in insurance sectors, right, maybe call transcripts. Those chatbot conversations are being used to train AI systems for fraud detection or customer support automation, although this is typically covered by terms of services sometimes. But then a lot of customers are unaware of the extent to which their data contributes to AI model training. So that needs to be looked at. Yeah.
Karen: Yeah, those are good examples. Do you know of any companies that ever made you aware upfront that they might use your data for training an AI or machine learning system? Or did you get surprised by finding out, maybe in their terms and conditions, that they were planning to use it? Or has that not ever affected you?
Kadri: Okay. For me, I can’t remember anyone telling me that I should consent to using my data for training AI. I can’t remember. Not at all, to be very honest. Not at all. I don’t know about you. Did you? Does anyone reach out?
Karen: Yeah, there was a fuss, not last summer but summer before, 2024, about LinkedIn informing us basically that they were going to retroactively claim the rights to use everything we put into the system up to that point, but we could choose to opt out from that point onward. I don’t know if you remember hearing about that. They did not exactly promote it. They just did it very quietly, and then some people raised a fuss about that, for obvious reasons.
Kadri: That makes sense. That’s the nice one. That’s a very nice one to do.
Karen: There actually were two places that you needed to go to opt out if you didn’t want them to use your information. And some people said, “You know, I want them to include my points of view in what they generate, so I’m okay with them using it.” And other people said, “No, I have my contacts, my networks, my work history, and everything else. This is all my information.” And they didn’t want it used for training. So I heard a lot of different opinions on that. I was just curious if, one, if you were aware of it, because they really didn’t promote it, and two, if you had any thoughts about that.
Kadri: Not at all. Not at all.
Karen: Yeah, they were not transparent about it, let’s just say.
Kadri: Wow.
Karen: You mentioned getting spam phone calls and such. Do you know of any times when companies using your data have caused any trouble for you? Privacy violations, loss of income, phishing, or anything like that?
Kadri: Yes, yes. I have a case. Yeah. I don’t want to mention their name, but there’s this startup, they’re very big in Nigeria, right? They’re a startup, like an alternative bank, right? But I’m always concerned because I didn’t set up on their platform. I never signed up. I’m very good with the commercial banks, right? But then I get calls every day about different promotions. Those calls are like an automated call where there’s an IVR [interactive voice response system] that set up prompts. Just the same call. Everyone is going to get the same call. And I’ve been behind the system. So sometimes I worry, is there someone that set up using my details? Why are they reaching out to me when I don’t even have an existing account, right? So I took it upon myself to email them. I’m sure they covered some things, but then after emailing them, they stopped, right? So obviously maybe they deleted an account or they just collected a lot of numbers and they’re not sure whether it’s a customer or something. So sometimes, I get a text, so I had to email them. I even scared them that we’d sue them. So at that point, I stopped getting those texts and calls.
Karen: Yeah, somehow one of my phone numbers got onto some sort of a list somewhere with the name of Valerie. I keep getting these texts saying, “Valerie, do you want...?” Okay, I’m not Valerie. And I don’t know if somebody made a typo when they typed their phone number in, and now my phone number’s forever stuck with some other name, or what. I just don’t answer or respond in any way to unrecognized numbers, and that helps. But it’s still annoying that they happen, because in some cases, just responding to those, saying “Hey, take me off your list” only serves to validate that, “Hey, this is a real number and there’s a real person here. So yeah, let’s send them some more stuff.”
Kadri: Wow. Wow.
Karen: For the works that you’ve created, things that you’ve written online, have you ever seen that content — I’ll just say plagiarized, whether by humans or by an AI tool?
Kadri: Yes. I’ve seen by humans, right? So it happened this person was my connection. I was writing about open banking. I can’t remember. So I saw this, I think a month after, I saw something like that written by a human being. But then it was just like, maybe plagiarize the beats, change the content, but there were still content from mine, right? At least they didn’t use the original context, but I could feel that that was me, anyways, whichever way. So I didn’t see as a threat. Maybe I’m inspiring someone.
Karen: I don’t really understand why people sometimes just copy someone else’s writing or something else that they’ve created and don’t give credit. You know, giving credit is easy, and I don’t understand why people don’t do it. But that’s maybe just me.
All right. Last question. We see that public distrust of AI and tech companies has been growing recently. What would you say is the most important thing that AI companies would need to do to earn and keep your trust? And do you have any specific ideas on how they could do that? Or do you think it’s not even possible to trust them?
Kadri: Yeah, so public distrust, outside even tech companies, while I was working in telecoms, I think we were the first provider of 4G when 4G LTE came around. I remember our customers didn’t trust us because we give you speed, but then their usage is higher because they’re buying time. It’s just a time for money. You open a YouTube, you’re able to watch it seamlessly, but then you’ve forgotten that there’s data somewhere. So a lot of people in that period thought that they were being scammed. This new technology that you buy 20 GB worth of data, but gigabytes worth of data is done in nanoseconds.
So we had to be transparent, communicate our intent and the service. You can even try to let down, maybe, your speed so that you’re able to do more, or we even recommend that you get more data. Because you’re the one that is actually using it. Yeah.
So the single most important thing that AI companies need to do to earn and keep trust is just be genuinely transparent. Be accountable about how these systems are built, trained, and used. This means moving beyond vague statements, and clearly explaining what data is used, where it comes from, how consent is obtained, and how risks I identify are managed. Transparency should be practical and understandable, not buried in legal-like language or maybe your privacy laws, and it’s somewhere very far, right? So we just have to be specific about what we’re using your data for. Provide meaningful opt-out mechanisms. And be specific about the limitations, the biases, and the failures, right?
So that’s it. Trust will not come from just marketing, but from consistent and very viable actions, over and over.
Karen: Do you know any companies that are doing this well right now? It’s okay if the answer is no.
Kadri: Yeah. I think. Back in Curacel, we are doing well, and currently even in Access Bank, we are doing well. Whatever we’re doing, the intention or maybe in activities that we try to engage, every phone call, we refer you to privacy. You get a link sharing “This is why we’re giving you a call”, because we’ve already mentioned that we are going to analyze the data now. So I think currently we do that very well.
That AI is not our primary solution, right? We use it more like a feature in-house. But when we do that, we do it with great intent not to complicate things. I also explain how your data is being used.
Karen: All right. Thank you so much for sharing all that. I really appreciate you coming on the show and sharing your ideas and experiences with us. Is there anything else that you would like to share with our audience? Anything that you were hoping I would ask you that I didn’t ask?
Kadri: Thank you for having me, Karen. It’s been nice. So for the audience, I have my GitHub for any of the projects that have been done. For those that want to collaborate on product data initiatives, you can connect with me on LinkedIn. I’m also leading a customer engagement, and I have a small community, right? Hopefully something can come up from that solution. We can go live this year. So anyone that wants to join the team or partner with me or even be a co-founder, they can reach out. Thank you.
Karen: Awesome. Thank you. For your customer engagement tool, what’s it called? And are you looking for beta users? Are you actively building, or where do you stand with that?
Kadri: So for the customer engagement, it’s a global tool. It’s not restricted to just Nigeria alone, Africa, Europe, America. Middle East – anywhere, right? So the goal of this tool is, it’s like a customer research tool, right? A lot of people collect data and don’t know what to do with the data, right? Maybe you are collecting feedback from customers. There’s an AI solution that’s maybe classified as complaints, send you triggers without even going into the platform, about maybe some things to watch out for. So it’s just more about automating engagement, how to hear from your customers. It’s been segmented, so there are lots. The more copies, the easier people can also check the idea. I’m open to feedback. I’m hoping to also go live this year, right? It’s being built currently.
Karen: That sounds wonderful. Well, good luck with it! Thank you so much again for sharing your thoughts on AI with our audience, and I’m looking forward to keeping up with you.
Kadri: Thank you very much, Karen. It’s been wonderful talking to you.
Interview References and Links
Adekunle (Mayowa) Kadri on LinkedIn
Adekunle Mayowa Kadri on Medium
Meet Adekunle Kadri (personal website; find details on Check AI and his current project here)
Kadri Adekunle Mayowa on Substack
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:
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.
Audio Sound Effect from Pixabay
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 Beth Spencer 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 😊)
