How To Become A Builder Not Just A User | A Conversation Featuring The Special Envoy on Technology, for Kenya

Guest: Ambassador Philip Thigo - The Special Envoy on Technology for Kenya

AI hype is loud, but the real question is quieter: are we building the infrastructure that makes AI work for us, or are we stuck as permanent users of other people’s systems? I’m joined by Ambassador Philip Thigo, the only African tech envoy, to unpack what “AI is infrastructure” means from a Kenyan and African perspective and why that framing changes everything from investment to regulation.

We break down the AI stack in plain language: compute, data, talent, use cases, and model innovation, including the hard truth that many African languages are missing from today’s dominant models. Philip argues that talent is the shortest path to sovereignty, particularly for countries that cannot realistically own massive compute or hyperscale datasets. We discuss small language models, local context, and why being a builder matters, with vivid examples such as flood prediction and the type of granular data that global models often overlook.

From there, we zoom out to the money and the power. Why do investors keep funding only “the model,” and what returns exist in energy, infrastructure, and applied AI use cases? How do development banks catch up when AI moves faster than traditional timelines? We also tackle data localization versus real data governance, the value of data, and the information battlefield of misinformation and disinformation, including the need to label synthetic content and raise the economic cost of harm.

If you care about Kenya’s AI strategy, Africa’s AI future, AI policy, data governance, and tech diplomacy, this conversation is for you. Subscribe, share this with a friend who’s building, and leave a review with your biggest takeaway.

Credits
Host: 

• Stella Gichuhi

Producer: 

• James Njoroge

Executive Producers:

• Harry Hare
• Agutu Dan

Transcript

Act 1: Introduction

Stella Gichuhi 0:00

What happens when technology meets diplomacy? On this week's episode, I sit with the only African tech envoy, that is Ambassador Philip Thigo, to talk about all things AI. Last year it was this is what AI can do. This year is we need to see what AI is doing. We talk about AI factories. There is a controversial Kenyan AI bill somewhere. Yes, I'll call it controversial. And most of all, we understand what leadership means and why we need leaders in this changing era of intelligence. Listen to the episode and let us know what you think. That's it from me, Stella Gichuhi. Good morning, Ambassador Philip Thigo. Good morning, Stella. How are you? I'm okay. Yeah? Yes. We see each other on the streets of tech. Now you're sat with me. Right. Let's talk tech and specifically AI. Got it. I didn't know much about AI until I heard you talking about AI, and I asked myself, in ni nini, what's this? But

AI is not new

Stella Gichuhi 1:06

you've taken the country along a very exciting journey the last couple of years. You have talked about AI, about skilling, about the need for the country to be recognized. You are honored and recognized at the UN level and guiding that AI conversation. But where we are now, to quote you, we need to talk. Let's talk about AI and infrastructure. Yeah. What does that mean from an African context? African perspective, sorry.

Amb. Philip Thigo 1:33

So I think part of the challenge around this discussion on artificial intelligence is that it's been led by a few countries.

Stella Gichuhi 1:44

Yeah.

Amb. Philip Thigo 1:44

And mostly it's China and the US. And then, of course, a lot of us have been trying to catch up. And so there has to be an understanding of AI beyond their articulation of what AI is. And so we have to understand AI from where we come from because they're really advanced in this conversation. And so I think for me, when I say AI is infrastructure, I try to bring people down to the basics of what we need to do so that we get to understand to articulate what AI is, but also we get to use and to build AI. And so I think if you look at how how I frame this, is in most cases is that we look at AI as a technology, but we don't understand the era within which we live in right now. Because AI is not new. No. AI has been there from the 1950s. But what is different at this time? And I think what is different at this time is that cumulatively, countries that have invested in infrastructure, in digital infrastructure, are the ones that are leading. Also, when I say AI is infrastructure, I talk about that that entire entire stream because everybody kind of dwells on the application, not even the application on charge GPT. Exactly, cloud or the models. And I don't think that's AI.

Stella Gichuhi 2:53

That's that it's it's a part of it, but it's not part of it.

Amb. Philip Thigo 2:56

It's not, and it's not useful.

Stella Gichuhi 2:58

No. What would you say about the rest of the continent, about the journeys? We're trying, but where would you say if you were to pull me to the side and say, Stella, I like what you're doing, but I need you to see more or less of.

Amb. Philip Thigo 3:14

Yes, again, and so I think part of also the sort of understanding has to be a shift in mindset.

Stella Gichuhi 3:20

Yeah.

Amb. Philip Thigo 3:21

Is that

The Fifth Industrial Revolution

Amb. Philip Thigo 3:22

I come again into the era we we live right now. So if you look at all the revolutions we've had, whether the first one was steam, second one was electricity, third one was automation, fourth was ICT and ubiquity of it with the fourth industrial revolution, the fifth one is intelligence, right? So you have to be in the moment. And this is something I tell people is that if if the if the currency of this era is intelligence, then that's what you need to be invested, investing in.

Stella Gichuhi 3:53

Okay.

Amb. Philip Thigo 3:53

And you cannot invest in every part of intelligence simply because of capability. And

Talent as sovereignty

Amb. Philip Thigo 3:59

so I take each and every country individually. So some countries, unfortunately, will have more capability than others. But for me, the shortest path towards participating in an intelligent era is knowledge. And so we need to build our knowledge base, and every country can do that. You don't have to have compute, you don't have to have even crazy electricity or energy, but you can build your knowledge base, and your knowledge base relies on your talent. And I think each African country has a talent base.

Stella Gichuhi 4:30

Yes, we do. Yeah.

Amb. Philip Thigo 4:31

And so and so I think for me, if I take your side, I'd say focus

The Shortest path to Sovereignty

Amb. Philip Thigo 4:34

on that. And talent is the shortest path to sovereignty. It is, because not every country can have compute, because not every country has a tremendous amount of energy. And we know only five or six countries can do this: Kenya, Ethiopia, South Africa, Egypt, Morocco, Nigeria, and slightly Senegal. Not every country can have data to that extent because data is what drives AI, but then it can the countries that are small, some countries are big, right? So not every country can necessarily have that huge capability of massive amounts of data sets. But you can build your own data capability, right? But not necessarily large. So I think people kind of go crazy around the large language models. You can also do a small language model. And so I think again, if I take people aside, is look at what you can do in your own capacity and capability and not try to be like another country. And there's no winning in this. And that's why I tell people that it's unlike the ICT age where everybody was running to digitize. Yeah. This is different. Okay. Because knowledge transacts very differently than than the digital era. Because knowledge can be exponential. You can be a small country, but you can be exponential, exponential in the era of AI. So if you just invest in that knowledge capacity, I think I think that's where people need to start.

Stella Gichuhi 5:58

Okay. So I need to work with what I have, build my knowledge base, and work just work with what we have. Work with what you have and stop competing. So I should not look at open AI and think, oh my gosh, no, I'm in Africa for Africa with Africa in mind. Thank you for that piece of advice.

Act 2: The Diplomatic Craft

Stella Gichuhi 6:13

You've touched on investment. Yeah. That's a term that's used in many spaces that I've been, and I'm sure you use that a lot, and those a lot of people have used it with you. Who is investing where and how can we unlock that investment?

Amb. Philip Thigo 6:31

Yeah. So I think if you look at the, and this is why we talk about the AI stack, right? So if you look at the AI stack, so AI stack is compute, data, talent, use cases of what we call, and then there's there's a fifth one we call model innovations. As in you have to in you have to invest in the models so that especially if you are from Africa where a lot of our languages are not in the current models. So we have to add that fifth one.

Stella Gichuhi 7:01

Yeah.

Amb. Philip Thigo 7:01

So I'd suggest you invest in the five. You can't invest only in one. No, you can't. Uh you have to invest in compute. You may not own it, but invest in compute in terms of using the compute. Invest in data in terms of building your data uh infrastructure and data architecture. Invest talent, I would not, I mean, that's where we really need, it's the hill we need to die on.

Stella Gichuhi 7:22

Yeah.

Amb. Philip Thigo 7:23

Because if you don't have talent, you can't use compute in your context. If you don't have talent, you cannot necessarily digitize. Because then data is is contextual.

Stella Gichuhi 7:35

Okay.

Amb. Philip Thigo 7:35

If you don't have talent, then you can't even do the use cases. Because then how do you apply on agriculture, on climate, on education? And then how do you train your own models? Especially in the current dominant kind of model context where a lot of the large language models are not representative of who we are. And so, how do you also have the same talent that then can build small language models that are more contextual and can solve our problems, but also understand our local capability and capacity models that use less power? Yeah. Models that can land on devices that are not necessarily, you know, complex or complicated. So I think for me it's the to invest in fact.

Stella Gichuhi 8:14

That investment pipeline. But where are the investments?

The AI Stack - Investment Framework

Stella Gichuhi 8:17

How do we compel an investor to say that's not our story and they should invest in us?

Amb. Philip Thigo 8:23

Because I think there's always, I think if you talk about investments, especially in artificial intelligence, again I say is people do not understand what AI is. Yeah. Because a lot of them are talking about investing in the models, in the frontier model companies, right? So it's it's open AI, it's anthropic, it's so for them, the bubble was busted when the when sort of the Eastern models showed you can do this at f by far less money. Yeah, yeah. Whether it's Quen or Deep Seek. And so I think that that's when investors started getting a little bit jittery. But then that's always been a Silicon Valley challenge. Okay. Where you're investing in these companies that have valuations that are so high in their initial phases that do not necessarily correspond to usual investments. And so always for me, when I talk to investors again, I tell them invest across the stack. I'm not saying invest on the frontier model companies. We're saying invest in infrastructure. There are returns on infra. Invest in energy. There are returns on energy.

Stella Gichuhi 9:21

Yeah.

Amb. Philip Thigo 9:21

Invest in talent. There are returns in talent. And if even those frontier model companies, you've seen are fighting for talent.

Stella Gichuhi 9:28

Yeah.

Amb. Philip Thigo 9:29

There are investments in the use cases. I think for me, that's where the real money is. Because when you apply AI on a dairy capability, then it exponentially increases output or productivity of dairy. That in most cases brings in more money. So potentially that's why I tell investors do not look do not look at AI as a software.

Stella Gichuhi 9:51

Yeah.

Amb. Philip Thigo 9:52

So AI is not software. It's not software. It's the stack.

Stella Gichuhi 9:55

Yeah.

Amb. Philip Thigo 9:56

It's an entire stack. And where the model, so a lot of them I say are in the model innovation, which is like the fifth stack. There are like four other pieces that are missing. And I'm not saying don't invest in the model innovations because that's where the frontier work happens, but it doesn't mean that's the only place you are invested in. So I think for me, if I ask, if you ask me where the investments are, I think the investments are looking at the wrong place to invest. So we need to re-educate the investment. Exactly. Investors need to understand what AI is and where the investment cases are and not just the frontier model companies.

Stella Gichuhi 10:27

I feel so validated because I remember telling an investor, if you don't invest in infrastructure, the pipeline you're required to invest in will not be there.

Amb. Philip Thigo 10:37

It won't be there. And remember the the constant, the only constant that has been there across the five industrial evolutions, and I always tell people, look at the five. If we didn't have steam, we'll not have electricity.

Stella Gichuhi 10:49

Yep.

Amb. Philip Thigo 10:49

If we didn't have electricity, we wouldn't have automation. Exactly. If we didn't have automation, we wouldn't have the fourth industrial evolution and you know, and IoT and Internet of Things and if we didn't have that, we wouldn't have AI. Exactly. So or this age of intelligence. I think and and and the first infra that was invested in still validated to date. Still lose electricity.

Stella Gichuhi 11:09

Okay.

Amb. Philip Thigo 11:10

Even for AI, it's the differentiator right now. Everybody's talking about energy. So it's the third industrial revolution, right? So I think for me, that's the investment case, not on software. Software will change.

Stella Gichuhi 11:24

And you can't treat in in my because I'm reskilling at the moment, and I can't treat an AI project the way I would treat a software project. I've I'm constantly reading, watching something, because it's very different. Now

Risk Appetite & Loan Timelines

Stella Gichuhi 11:36

let's take this upper notch. There's World Bank, there's AFTB, there's UNDP. What's their appetite been like around their investment?

Amb. Philip Thigo 11:46

It's quite an interesting conversation. I think it it depends on who they are, just in terms of their personal right. So UNDP, by the very nature, or being a development organization, their thinking has always been so how do you bring those who are in the margins or who do not have access to begin to access this new technology? So that's the mindset. So get those who are in the margins to access AI and use it to solve their problems, but also see how this can uplift people from poverty or sort of move the needle around development solar as your NDP. I think if you look at the FDB and the World Banks, I think for them being traditionally banks, my sense is that they've come late. Oh. Yes, they've come late in the game. Also because of just the kind of legacy mindset, risk, appetite, and risk portfolio.

Stella Gichuhi 12:38

Okay.

Amb. Philip Thigo 12:38

Because if I tell you invest in compute, a development bank will not understand what that is. And we've had these discussions before. You've seen only, I think, a week ago or a month ago is when you see them start to move.

Stella Gichuhi 12:50

Yeah.

Amb. Philip Thigo 12:51

With FDB looking at this 10 million, $10 billion AI fund.

Stella Gichuhi 12:56

Yeah.

Amb. Philip Thigo 12:56

And even then, if you think about it in the grand scheme of things, it's nothing. Nothing. $10 billion, it's nothing. Compared to what OpenAI is raising or what Ian Lekan just raised last week, or what Mira just raised, Mira Murati, the former Open AI CEO, you can actually see it destroyed. Yes, I've seen that. Yeah, to the extent of how they build funding or resources for AI.

"The Rule of This Revolution"

Stella Gichuhi 13:23

So reimagining, re-educating, and that will need a whole rethink on their timelines because the loan disbursement lifetime li timelines vis-a-vis the development of AI, I mean, technology is outpacing. It does.

Amb. Philip Thigo 13:39

And the point of this, and and and that's I mean, if if if if you've heard me talk, I keep on saying that the rules of this revolution are not the same as the previous ones. And we seem to have come from, you know, from the previous ones into this one with the same rules. And when I say this one, it's the last 15 years. And so I think this rapid pace of change has challenged how how we think or how we even reimagine the rules on everything, whether it's whether it's I mean, and and we've seen it, right?

Stella Gichuhi 14:11

Yeah.

Amb. Philip Thigo 14:11

Whether it's it's even understanding what computers, if even in the understanding of it, even in the governance of the entire I stack, um rules just don't apply. Rules just don't apply, guys.

Private Sector vs The Human Side of AI Change

Stella Gichuhi 14:23

Where's the private sector and all this?

Amb. Philip Thigo 14:25

Private sector, again, uh, it depends. It depends, it depends on the sector of the private sector. So I think of the private sector, right? So if you look at the tech the tech sector within the private sector, they seem to have it all figured out. Uh and even then and even then for them, it's it's very much around trying to reinvent themselves, right? So you've seen, for example, the big tech companies are no longer big tech in AI. You have it's it's really the fabs and the and the labs and the frontier model companies that really are leading in this, right? So you have one open AI that almost now is the one powering Microsoft. So until a few days ago, I think they're rethinking the agreement. It's been, you know, it's been cloud and anthropic powering even the military. Yes. Now we know.

Stella Gichuhi 15:17

Is it anthropic powering?

Amb. Philip Thigo 15:18

Anthropic, yes. Oh, yes, yes. Now we know, right? From the Pentagon challenges that we saw. Yeah. But I think we've seen AI kind of rebalancing the power even within the tech sector. So that it's the frontier model companies that are now the big tech in my view.

Stella Gichuhi 15:36

Yeah.

Amb. Philip Thigo 15:37

If you go into now your traditional private sector companies, then you see a lot of them are being a struggle.

Stella Gichuhi 15:42

Yeah.

Amb. Philip Thigo 15:42

So the big manufacturers are struggling. All of a sudden, you know, small manufacturers are doing rapid manufacturing, just leveraging on AI, and they're outfacing big manufacturers. Logistics, we've seen a lot of them are being challenged. Folks in geospatial art observations, we are hearing GO AI now. A lot of them are being challenged. A lot of them who had legacy infrastructure, the big software privacy companies, right? Yeah, a lot of those guys are struggling. Yeah, coders. Yeah. You know, if if you are a company that depends on coders, again, we assume you're struggling. If you are in first FMCG, sorry, people are struggling. You know, it's simply because uh a little logistics company that now has AI embedded by enabling them because they can quickly deploy AI. Yeah. Uh has made the big kind of those your your big traditional sectors, media.

Stella Gichuhi 16:33

Yeah, yeah.

Amb. Philip Thigo 16:34

Media are the worst hit. You know, it's to the extent of of we've seen like lead influencers or independent media companies now online leveraging on AI to produce amazing shows that are now captive.

Stella Gichuhi 16:51

And music.

Amb. Philip Thigo 16:52

Music, yeah.

Stella Gichuhi 16:53

I had I had a family member create a very beautiful song for me, and I said, Is this AI? And said, Yeah. Exactly.

Amb. Philip Thigo 16:58

Okay, also so creative and creative economy and creative industry, right? A lot of them, and we saw the Hollywood strike two years ago. Yes, yes, and issues around IP and a lot of that. So I think private secretary in most cases, I think, have been most challenged. One is because I and I think it's both ways, right? So one is because you can do a lot with little, with a few people, so you have to lay off folks. But also there's a part where you have to keep up because then technology keeps on changing, then the industry is changing, things are moving very quickly. We've seen airlines and travel agents. Yeah. You know, people have agents now booking tickets and then sort of things. Yeah. Well, it's early stage. It's early stage, but quite. But then the private sector is being challenged. Yeah, yeah. Um, we've seen issues. So I think for me, if there's a if if if there's one line that I

Technological Duality

Amb. Philip Thigo 17:47

think if you ask me why is private sector, I would say a lot of them are being disrupted.

Stella Gichuhi 17:54

Right, so they're going through their own disruptions.

Amb. Philip Thigo 17:56

Yes. Revenues are shrinking, compressing. If you look at media, for example, right? Yeah. So revenues have gone down. Yet other independent media's revenues are increasing. Yeah. So there's a question private sector has have to ask themselves. How come somebody else's revenue are going up while mine are shrinking? And and for me, the differentiator is it's that person plus AI. That's the differentiator. That that person has been able to harness AI.

Stella Gichuhi 18:25

Yeah.

Amb. Philip Thigo 18:25

And integrate it into their business process so that they are now getting exponential.

Stella Gichuhi 18:29

So the next phrase, what I I came across this new term, a technological duality. Now with AI, it's you win some, you lose some, but at an accelerated rate.

Kenya's Position Between US and China

Stella Gichuhi 18:39

Right. Geopolitically, about is it a few days ago you were with the Amchand Digital Economy Task Force. Tell us, what was what was that experience like? What are you how are you positioning Kenya in such a task force? Because I mean it's gonna go straight to straight statehouse, I'm assuming. White House, not state house. Sorry.

Amb. Philip Thigo 18:58

And I think the the the Kenyan posture has has been very clear, right? I think it's to recognize that artificial intelligence right now is really between or the capabilities lie between two countries. Yeah. It's the US and China. It's very clear. Yes. And so if you look at folks who have developed and built the stack, and so I think our position has always been that for us as Kenya, it's not to choose either out. What is it? We we we don't put ourselves in a way that we have to choose. We don't. First of all, I think especially when I talk about sovereignty. Yeah, yeah. Sovereignty is not about you choosing other people's identities. It's about it's it's it's it's about what matters for Kenya. And I think what matters for Kenya is a best of class. And so, and and as a country, we know who we are. I mean, we're not struggling with our identity. No, we're not. And so I think for us, the idea then of having a conversation with the with with the digital economy task force is actually it was more or less of a consolidation of of conversations we've been having with the specific companies. That is Google, Microsoft, Anthropic, Open AI. And it's to recognize, which is true, is that US firms are the best of class on on AI. And I think for us it was more or less so how do we consolidate this? And it and it was and it's not whether we partner, it's it's how.

Stella Gichuhi 20:23

Okay.

Amb. Philip Thigo 20:24

And I think for me, the and and people I think kind of confuse that. And I think for us, and the idea of having a chat with that is that finally we've come to a conversation where the US recognizes Kenya is a and so the how we're going to partner is how can we get the best of class of the entire AI stack from the US companies, right? So it's it's cloud computing, it's compute, it's it's research and development capability and capacity, it's it's talent. We have a lot of Kenyans in in America companies. It's it's also use cases. There are some things we share. For example, healthcare, okay, uh food. So there's some things we're already partnering between Microsoft, Microsoft AI for good, ourselves, Carlo and Ministry of Agriculture, around how we can leverage AI to improve food systems and food systems innovation. So I think it was very much around can we consolidate this now? Because we've been doing it with individual companies, so that we can get the best of class of the entire staff. Got it, got it. So that you know, and and so how do we kind of take our partnership to the next level around this new intelligent economy? I think we've never had this

Builders, not Users

Amb. Philip Thigo 21:33

conversation.

Stella Gichuhi 21:33

And and what what would your dream look like? If you could look into a crystal ball and this work that you're doing and say, I'll be so happy and proud of our work when X happens, what's that X?

Amb. Philip Thigo 21:46

I think that this conversation for me is is if if we become builders and not users of technology, because yes, we say AI right now is being led by a few countries and a few companies. Yeah. But being led for us is fine. Okay. But we but for us, we are Kenya, we run marathons. I think it's to the extent for us, is that are we able to build?

Stella Gichuhi 22:12

Yeah.

Amb. Philip Thigo 22:13

Because as you said, those models are fine, but they're not contextual. Our challenges are different. I always say if you were to train a self-driving car in Nairobi, it would work in San Francisco.

Stella Gichuhi 22:26

Actually, it would even work a lot better in San Francisco. Exactly.

Amb. Philip Thigo 22:29

If you were to bring the San Francisco way more into Nairobi, the car would not move. You know, and so and so we have something to offer. Yeah. Uh to also the US, in my view, to the extent that we have a lot of capability around training, we don't have legacy challenges, we have a lot of use cases. We have a lot of problems. And so having a lot of problems means we have a lot of use cases. Take a point of case of the floods we had. Yes. On Saturday, right? So I could I saw a lot of AI applications. Lots of it.

Stella Gichuhi 23:00

There's a gentleman who uses AI to predict the outcome and predict the floods and floods.

Amb. Philip Thigo 23:06

Because a lot of them can do floods. They can do flash floods.

Stella Gichuhi 23:08

Yes.

Amb. Philip Thigo 23:09

Right? So flash floods means you can go very specific. I can tell you, for example, around outside uh what is this place called? Integrity House, Integrity Center. Yeah.

Stella Gichuhi 23:21

Billy Mali Road.

Amb. Philip Thigo 23:22

You couldn't pass there. Nope. But you see, ideally, other models not tell you that. But if you do a local model that has has has, for example, done uh like uh like like the model has looked at the data sets around the elevation, the slopes, the constructions, you know, the lack of green spaces, you know, the drainage and historical data. I would ordinarily tell you that it's not just floods, it's the velocity of the water that is running right up from State House Road coming down to integrity house that makes that place very dangerous.

Stella Gichuhi 23:59

Yeah.

Amb. Philip Thigo 24:00

Which would have made Heron Court notebook clients at night. Because none of their clients would have no, they didn't sure. That's what I'm saying. They couldn't they couldn't tell you. It was impossible. It was impossible, right? And so and so I think I think for me that's that's if you ask me, that's the one thing that hopefully we can do is yes, engage and partner, but then with the best of class, but then be able to build your own contextual

AI Fluency as Negotiating Power

Amb. Philip Thigo 24:24

models. We need to invest in knowledge.

Stella Gichuhi 24:27

When you say knowledge, do you mean reading what would please help? Everything, just your fluency.

Amb. Philip Thigo 24:33

AI fluency, okay. Fluency, not even the stack. Okay. The entire thing. If you're going to negotiate on compute, you better understand what you're doing. You better know what you're negotiating, right? Because if you're negotiating on compute only thinking it's the GPUs, you're lost. You sound like Michael. Compute is not just GPUs, right? So if you send folks to negotiate and they are not fluent even in the language of negotiation or what they're negotiating on, then you will get a bad deal.

Stella Gichuhi 25:00

Yeah.

Amb. Philip Thigo 25:00

And then you'll end up blaming a company that was so happy that the folks they were talking to had no clue what they were working on, right? So

"no country or person"

Amb. Philip Thigo 25:07

again, I come back to that piece, which is why I tell every country at least needs to invest in knowledge. And I think the president says it very well. Because Zakoti told us that no country or person can develop beyond their knowledge capacity. True. You just can't. You're limited. And so do not necessarily blame people for your limitations.

Stella Gichuhi 25:28

Thank you.

Amb. Philip Thigo 25:28

And so, how do you continuously update, skill, reskill, upskill, improve your fluency around these terminologies, but also your basic and common understanding. So that when you're negotiating, you're negotiating from the understanding of the subject matter, but also understanding of your own context. Yeah. Because otherwise, sometimes you end up with a white elephant.

Act 3: The Strategy

Stella Gichuhi 25:52

Yeah, that brings me to my next question. It's a hot one. Feel free not to answer, but I'd love for you to answer. There's a draft AI bill floating somewhere. And when you read that bill, I put it down after I read an AI commissioner or a regional AI commissioner. I don't know. There was languages in there. Has that bill been drafted with knowledgeable people, or is it just being put out there for the sake of being put out there? Or

The AI Bill Draft

Stella Gichuhi 26:19

let me be diplomatic and say we are in the preliminary stages. Therefore, there needs to be a lot more conversation going into it. Whilst I'm very tempted to attack, I also need to be diplomatic.

Amb. Philip Thigo 26:32

And I'll say honestly, I I don't know where that bill came from. Honestly, I would say it's not part of our strategy. No, it's not. I just answered that.

Stella Gichuhi 26:41

Because it goes against everything you talk about.

Amb. Philip Thigo 26:45

I mean, we have a we have a timeline. Yeah. We started with the AI principles last year. We went into doing an AI strategy in the middle of an AI policy.

Stella Gichuhi 26:54

Yeah.

Amb. Philip Thigo 26:54

So a bill was not part of our yeah, I've had your conversations.

Stella Gichuhi 26:59

You've never you you've always been pro-maturity, exactly innovation, build your talent.

Amb. Philip Thigo 27:04

But also it's not yet time to do um a hard law.

Stella Gichuhi 27:09

Yeah.

Amb. Philip Thigo 27:10

Because we said we want to test a lot of these things. Yes. Because AI, as I said, is not like any other technology. And we don't think also one bill covers AI.

Stella Gichuhi 27:21

No.

Amb. Philip Thigo 27:21

Because as you said, if you look at the stack, right? There's talent, which is education, which means you still have to reform the education stack. Exactly. There's energy and power. It means you have to there's investments, there's taxes in terms of the components of the AI stack.

Stella Gichuhi 27:36

Yeah.

Amb. Philip Thigo 27:37

You know, AI touches every aspect of society, then I seriously don't think one law covers it. Covers all covers it. We still need to evaluate, and that's why we are going through the policy process for that and through public participation. It's been really engaging, including all sectors, which is still ongoing so that we can understand the AI opportunities and risks.

Stella Gichuhi 27:58

Yeah.

Amb. Philip Thigo 27:58

You can't just all of a sudden come up with a bill because of a risk. That's true. You know, this is not, it's not how we, it's it's not who we are.

Stella Gichuhi 28:08

So there needs to be given more time into that bill than whoever anyway. Okay,

Data Localization: Protective vs. Protectionist

Stella Gichuhi 28:12

fine. Data localization is part of that stack, right? Yes. Is it protective or is it protectionist?

Amb. Philip Thigo 28:22

Well, I think if if we talk about data localization. Well, if you look at how we look at our data, right, and this part of the data protection legislation, right? It just says you need a copy. You need a copy. You you at least need a copy of that data in country. Yeah. And that should be fine. So I don't think localization is a challenge, in my view, in the context of the current law. It's not prohibited. Okay. To the extent of how we understand or manage our data. I I I honestly think we need to move beyond protection and localization to management of data. Because the the challenge right now around building models is not about data protection. We don't even know how to manage our data.

Stella Gichuhi 29:09

So we need a data act? Not really.

Amb. Philip Thigo 29:12

We actually don't. I think for me it's it's it's more, and and we talk about this issue around data values. Yeah. We need to figure out how to how to create a value around

Value from Data

Amb. Philip Thigo 29:24

our data. So if you were in the airport on Saturday at 11 pm, it didn't rain. But if you are to get to Parkside, which is where all the rain is, it started to rain there. The minute you get to Nyay Stadium, it was poor.

Stella Gichuhi 29:42

Yeah, yeah.

Amb. Philip Thigo 29:43

So think about it. Yeah, yeah. So think about it, right? So but think about how many base stations were there in between the airport and those base stations. Base station is a cell tower.

Stella Gichuhi 29:57

Oh, okay.

Amb. Philip Thigo 29:58

The cell tower collects weather data. And it collects weather data for operation. Because when when when there's a lot of rain, it can boost automatically boost the signal. So it's just efficiency. But it's not for any. So imagine that that's a Faricom or air tower, base station could have shared that data to make that weather data available at a very granular level. And it would cost them nothing because it's for operations. So that's the social value of that data. Then there's economic value of the data, which means that data has business sense for that company, as in they make money out of that data. Safari commallet.

Stella Gichuhi 30:44

Yep.

Amb. Philip Thigo 30:45

Yeah. Right? So they use that.

Stella Gichuhi 30:47

Yeah.

Amb. Philip Thigo 30:47

And then there's the developmental value of the data, which means you spend money to develop that data. Got it. So if you do the three, it means you can actually cost the data. Which then makes it easier to share that data because you understand the value. Because then you're sharing that data, either you put it behind a paywall because it's costing you.

Stella Gichuhi 31:08

Yeah.

Amb. Philip Thigo 31:09

Or it can be for public good, which means it can be released for free. Yeah. That it can help the public or public value. And we've seen it. I mean, Rescue, which is an ambulance company here in Kenya, built their entire practice on top of Uber data.

Stella Gichuhi 31:27

I didn't know that.

Amb. Philip Thigo 31:28

Uber mobility data. Uber release that data.

Stella Gichuhi 31:30

Oh, yes, okay.

Amb. Philip Thigo 31:31

Rescue now can can avail an ambulance instantly, and the response time is 15 minutes from 169 minutes. And so it means, and that's what I'm saying, it means we need to have a conversation on the data value, on the value of data. In that way, you see you're open to share the data because you understand the data.

Stella Gichuhi 31:53

You understand the value. But you've been you've been championing that conversation for a while, if I remember correctly. Yeah. What's not happening? What what are we not doing as the people?

Amb. Philip Thigo 32:04

It needs to happen at scale, right? It can't be just, you know, real conversation. That's why we're doing the white paper. Okay. So that we can get folks. And and for me, that that is why I said for Kenya it needs to be a different conversation. It needs to be that discussion. Because that's what everybody has a problem with. That you want me to give to give me data, but then I spent money building it. Yeah. There's IP issues involved. So then let's have a real conversation on the value of data.

Stella Gichuhi 32:28

Yeah.

Amb. Philip Thigo 32:28

Right? And so if everybody understands, then it means we can build a platform where we where data can be exchanged based on its value.

Stella Gichuhi 32:37

Very simple. Very simple, but like anyway, so there's there's there has to be a start. There has to be a start.

Information Warfare

Stella Gichuhi 32:46

Ah the information warfare, the information battlefield. We've talked about the good side of AI, the data, who's not doing what who should be doing it, but things are about to go south if they haven't already in certain worlds you sit in. What do we need to do differently?

Amb. Philip Thigo 33:03

I think Yeah. Again, let's go back to knowledge and I'll keep on circling back into the knowledge. And I like that you're doing that. Yeah, you have to go back into the knowledge, right? So part of part of you understanding information, misinformation, disinformation is first of all you understand what isn't what is real, what is not. Right? And so if you don't have the knowledge to discern the reality or the lack of a shared reality, then there's a problem. I

Platform Responsibility for Synthetic Labelling

Amb. Philip Thigo 33:34

think one is that there's a responsibility. Responsibility, and this is something I think in our role as in the UNAI advisory body, we were very clear with the big tech companies that the only people who will know from the onset that that information is not real is the company within which that information is being produced. Okay. And so the whole idea is can can there be a policy, even from the company, that that that information is labeled, synthetic or fake, yeah, or manipulated, or whatever. So and and they have a way to do this. So we've seen TikTok as you do that.

Stella Gichuhi 34:13

Yes, yes.

Amb. Philip Thigo 34:14

And and so I think it starts right from there. And this is why we say it's not about stifling creativity, it's that people being very open to label that information.

Stella Gichuhi 34:24

Yeah.

Amb. Philip Thigo 34:25

Uh so that the consumer knows that that information is real or not real.

Stella Gichuhi 34:31

The the big companies are doing it, but what about the African companies or you know, that individual who wants to run a very dirty disinformation campaign because now that's not a company, that's an individual who's out to but still run it on a platform.

Amb. Philip Thigo 34:46

Ah that's important, right? So so so I think the platform has a responsibility. And we've seen again, I I go back to TikTok. We were the TikTok's safety summit last week. Yes, yes. And we've seen TikTok has come a long way.

Merchant of Mayhem

Stella Gichuhi 34:59

Yeah, I'm you can only a company can only do so much, you know, the individual. And what what do we need to do to that individual who I call them a merchant of mayhem?

Amb. Philip Thigo 35:09

That's to be a cost.

Stella Gichuhi 35:10

And and I think that's the question I should be asking, yeah.

Amb. Philip Thigo 35:13

Where we where we I think are not being very clear is that every piece of misinformation, disinformation or defect has an economic incentive. And I have a sense that a lot of the punitive measures do not look at the economics.

Stella Gichuhi 35:32

Okay.

Amb. Philip Thigo 35:34

They look at the penal nature of it, right? So that oh you'll be jailed for 20 years. Yeah. Or pay a fine, X. You know, I think we must make the cost of misinformation very expensive. Right. So that because that is an incentive. Okay. Even if you see post that is of a political nature, yeah, it is still being underpinned by an economic incentive. Nobody just does a political post just because they hate the other candidate. There's always money driving that. And and that's a proven, it's a it's a proven study. So you must make the cost very high. I would I would put a serious fine.

Stella Gichuhi 36:13

So it's like ten a thousand dollars for 10x.

Amb. Philip Thigo 36:18

I mean, there has to be a way we can calculate the cost because of the irreparable harm to society. Yeah. Irreparable harm to reputations and family, but also breach to democracy, rule of law, human rights, you know. And so I think unless we make it expensive. It's it's less than companies.

Stella Gichuhi 36:38

Yeah, companies can afford it, but to the individuals, I think. Yeah.

Amb. Philip Thigo 36:45

And and and again, as you say, the the goal is not to control information. Yeah, the goal is about information integrity and preservation of societal norms and people's individual rights and freedoms. I think I think there has to be a cost.

Leadership as the Differentiator

Stella Gichuhi 36:58

Knowledge, knowledge, knowledge, knowledge from the onset. It's been knowledge, knowledge, knowledge. Do you sometimes feel like you can just go in and whip everybody to to act? It's annoying. But yeah, we've been waiting for it. Do you just look at the things, guys? We're in the age of the internet, you can research this, but they just refuse to, or competing priorities as well. Could that be a deterrent? Actually, no.

Amb. Philip Thigo 37:21

I think it's leadership. Oh, okay. I think it's leadership. I again, uh, if you look at if you look at companies that are exiling in this era, it has to do with leadership. It has to have done with how a CEO or our board or our cabinet or president has been able to move a little bit differently. Understanding that that this era requires kind of a different approach on how you either make investments, how you make decisions, your risk appetite, your innovativeness, where you invest, how you champion specific aspects, right? So I think in most cases it's that, it's leadership. Secondly, again, I think Hollywood has done a number of a lot of us, which is what I say. It's trust it's a trust issue, right? So we we we we live in an era of diminished trust. So people are less trusting, right? Okay, including untrust and lack of trust, you know, builds fear. So a lot of folks spending a lot of time living in their fear. So when I say we seem to have a challenge of imagination, is actually because people are afraid. And you're afraid of oh AI will take over the world. Okay, we will lose our jobs, okay. But then this is not the only revolution that has come with job losses.

Stella Gichuhi 38:42

Yeah.

Amb. Philip Thigo 38:42

It's not.

Stella Gichuhi 38:43

So go back to history.

Amb. Philip Thigo 38:45

Yes, I mean it's not. When the printing press came, it doesn't mean journalists lost jobs.

Stella Gichuhi 38:51

No, they pivoted, reskilled, or just left entirely and went elsewhere.

Amb. Philip Thigo 38:56

Were able to come into the printing press and do mass newspapers.

Stella Gichuhi 38:59

Yeah. Yeah. There's an opportunity for everybody.

Amb. Philip Thigo 39:03

Again, this era, like any other era, will have winners and losers. And so whoever will be a winner in that particular inverted quotes is who has been able to understand the the uses of it. Yeah. The misuse of it. Yeah. And then the missed use. What is the what and the misuse for me is the biggest one, right? If I do not use AI, what is the cost of not using it? Versus use, misuse, missed use. Okay. The third one is always the worst. The opportunity cost. You know, the cost of inaction. And people don't quantify that. Yeah. And that is where people are right now. Oh, we lose jobs. So you are so you're not acting? And you'll just wait for jobs to be lost.

Stella Gichuhi 40:00

No, no, get off the internet for the stuff. But that's what everybody's doing, right? Download books, PDFs. Just investing in capacity. It's gonna be tough.

Amb. Philip Thigo 40:06

Invest in capacity, invest in knowledge, invest in infra, invest in just in invest in just understanding the trajectory of this technology.

Stella Gichuhi 40:13

Yes, I agree.

Amb. Philip Thigo 40:14

With your vertical, right? So where will you as a media house be in this trajectory of this technology and what do you need to do very quickly? What do you need to do? To to pivot. But yeah, but you can't just not act. No. You know, and that is what I see with a lot of our companies

Navigate Uncertainty

Amb. Philip Thigo 40:35

here.

Stella Gichuhi 40:36

Yeah, adapt or die, pretty much. When agile came out a few years ago in the UK, it was change or die. Now it's adapt or change. Exactly. Yeah.

Amb. Philip Thigo 40:45

So so I think there's a again, it points down to leadership.

Stella Gichuhi 40:48

Yeah.

Amb. Philip Thigo 40:49

But because who has to make those decisions? It's a leader who has to make those decisions. So if the CEOs don't do it, if the CFOs don't do it, the CIOs don't do it, then who's gonna do it? Um and and and so we I I have a tendency of not understanding why people people blame AI. If it wasn't AI, it would have been something else. It was COVID in in five five years ago. It was the elections of the exactly so it was it. So we blame COVID. So everybody would just, you know, well, when people are going online and you're just sitting around waiting for COVID to end. So so there they'll always be, I always tell folks that we live in an era of uncertainty, and people have to be comfortable with uncertainty. No uh because something will happen. Right now, there's a war in the Middle East, yeah, something will happen. There will always be something coming at global scale that will challenge businesses or governments or countries or individuals, and that's the era we live in right now. Yeah, there's just a lot of uncertainty. But so will you not act because of uncertainty, regardless of which whether it's a tech thing, whether it's a pandemic thing, whether it's a health thing, whether it's a food thing, whether it's a weather thing.

Stella Gichuhi 42:07

Yeah.

Amb. Philip Thigo 42:08

I mean, we just got floods in Nairobi, Nairobi City. Or a food thing. Exactly, a food thing. So how do you how do you have leadership that can navigate? It's about navigation. It's not about it's not about certainty. No. It's about navigation. It's how will you it's navigation and resilience, right? So how will you build your resiliency? But also you're in a navigation capacity so that you can navigate your company, your business, your country, your yourself. Your mental status. Yourself through through uncertainty. And and we've seen folks who who are who are navigating this are the ones who are flourishing.

Stella Gichuhi 42:47

Okay. So you're flourishing.

Amb. Philip Thigo 42:49

Yes, I am. Unapologetically flourishing. I'm one of them.

Stella Gichuhi 42:54

I love this. I want to wrap this up, but before

Techplomacy: Multilateral to Multi-Stakeholder

Stella Gichuhi 42:57

we wrap this up, you have been very vocal as well. Gosh, you're vocal about a lot of things. I do, you've taught us a lot. Techplomacy. Techlomacy. Yep. I've you I've had the pleasure of sitting in one of the many forums you've created around the Techplomacy Connect, you call it. What are you hoping to achieve? What's your again, what's your wow, I did that, guys?

Amb. Philip Thigo 43:26

Well, it's not about me.

Stella Gichuhi 43:27

It's not about you, but sometimes you just I have to say, I did that. I had a vision, I've executed it. This is the diplomacy agenda.

Amb. Philip Thigo 43:35

I think if you look at diplomacy and and the idea behind diplomacy is is understanding that technology has become integrated and integral to diplomacy.

Stella Gichuhi 43:44

Yeah.

Amb. Philip Thigo 43:45

And not diplomacy in that I mean just I'm just an ambassador, so potentially this is how things align. That's how God works. But I think there's definitely a conversation that technology companies have become almost as Largest countries. And the current multilateral system does not recognize that. Again, let's go back to the rules have changed. Yeah. And the rules of the previous revolutions do not apply. So you have five trillion dollar companies that are now running big infrastructures. Yeah. That even are telling member states, even if you have the money, I will not sell you. Right? So, and or you have a company that is a three three trillion dollar company that has 15 million of your citizens on that platform. Either monetizing or engaging. Are we talking about ChatGPT? No, this is uh this is uh this is meta. This is all this is meta. TikTok is 17 million Kenyans. Uh Kenyans alone, Kenyans alone, Chat GPT, as you know, Kenyans are the number one users of Chat GPT, but for emotional advice or for companionship, right? So you have you have companies that are captive of your citizens, captive little of your citizens, and so you think you will not engage with them, you know, and so that's where diplomacy comes in, is understanding that companies are so big right now that they are almost akin to countries or multilateral organizations, and so you have to engage with them as if you're engaging with your previous counterparts, right? So and so that engagement means that there has to be a dialogue on how you govern. Because as I said, you are almost 15 or 17 million of your citizens in that or spending more time on that company's platform than even other things in your country, like Kenya, four hours and 13 minutes every day. Kenya spend on social media more than any other country in the world. How will you not engage with those social media companies? Because again, your citizens are uh captivated and are spending a lot of time on those platforms, and those platforms shape opinions, they shape identity, they do, shape behavior, how we shape behavior. And so I think for me that's diplomacy, right? So so then it means how do you then navigate those subjects, whether it's misinformation, disinformation, whether it's climate, whether it's rule of law, whether it's democracy, whether it's businesses, whether it's creative economy, that you need to engage with those companies and bring everybody on board. So it's so breaking your traditional multilateral member state to member state to multi-stakeholder. Okay. So I think for me that's diplomacy, and hopefully, as you said, what we hope to achieve, I think, is people to understand that that's the shift.

Stella Gichuhi 46:48

Right.

Amb. Philip Thigo 46:48

The shift is from multilateralism to multi-stakeholder.

Stella Gichuhi 46:52

Yeah.

Amb. Philip Thigo 46:52

And and and and multi-stakeholder in a way that you you reinforce the rules-based system. Okay. Because everybody now does rules.

Stella Gichuhi 47:02

Yeah.

Amb. Philip Thigo 47:03

The platforms have their community rules. We need to have them. But they have the community rules which are stronger than your national law sometimes. And sometimes there's a there's a clash between your national laws and your community rules. Um, and so how do you create an alignment? And that's why you see me engaging a lot of these companies, is just to align.

Stella Gichuhi 47:25

National laws are not a lot of things.

Amb. Philip Thigo 47:26

Where there's uh where the two are not saying different things, so that then it leads to confrontation. Got it, got it. Uh so we I call it um going beyond the adversarial democracies. Democracy should not be adversarial because each the platform says they are democratic. Yeah, Kenya says it's democratic.

Stella Gichuhi 47:45

Yeah.

Amb. Philip Thigo 47:46

So should we be careful so that our democracies are not adversarial? We have to recognize that particular aspect. And I again I keep on saying rules have changed. And so, how do we get our session? And that's why you see for Kenya, unlike other countries, we don't have a clash. No, we have a we have a platform to engage, and so every time there's a there's potentially a flashpoint, we will see me engaging with uh tech companies, and so you'll never see an extremity of a shutdown, or uh, you know, you you will never see

Conclusion

Amb. Philip Thigo 48:24

it.

Stella Gichuhi 48:24

And I I want to top up, we have you because we're lucky, we're lucky. Again, leadership, right? Yes, we do we have to be president, right? So they need to have your positions, you're the one in that position. Yeah, same thing, yeah, yeah, yep, yep. And you are the first and only one on the continent here, which is a sad thing, but uh well, well, it's not sad for us. I guess the rest of the countries need to get themselves an ambassador Philip Thiggle. Well, thank you. Somebody else, their own version of you, okay. Well, thank you so much for all the knowledge I've learned quite a bit, and thank you for honoring our invite for our listeners. That has been Ambassador Philip Thiggle, the first and only tech envoy from Africa.