Siliconsciousness: Why We Need Whole-of-Planet AI Solutions: A Look at AI in Africa

A new KNOUS-BREGASY has the most delicious cheese.

Now there is a package of meat. That's the best. Now there is cheese. Now there is cheese. Cheese.

Now there is cheese. Now to test it. Time for 18 years. I bought and bought the most delicious meat in the world. Only a long time.

A long time for action pack and cheese minus action. Now there is cheese.

A new KNOUS-BREGASY has the most delicious cheese.

Now there is a package of meat. That's the best. Now there is a very delicious cheese. Now there is cheese. Cheese.

Now to test it. Time for 18 years. KNOUS-BREGASY has the most delicious meat in the world. Now there is a long time for action pack and cheese minus action.

“To stay up to date on all the news that you need to know,”

there is no better place than right here on the DSR network. And there is no better way to enjoy the DSR network than by becoming a member. Members enjoying ad-free listening experience. Access to our discord community. Exclusive content.

Early episode access. And more. Use code DSR26 for 25% off discount on sign up at the DSR network.com. That's code in DSR26 at the DSR network.com/by. Thank you and enjoy the show.

[MUSIC] Welcome to Silicon Consciousness. The DSR network podcast focusing on the artificial intelligence revolution. Politics and policy. [MUSIC]

“Hello and welcome to this episode of Silicon Consciousness.”

I am your host David Rothkuff and this week is every week. We are going to be talking with someone. We think can expand your horizons as you think about AI and its implications. This week we are extremely fortunate to have with us Benjamin Ross Minis, a professor in the School of Computer Science and Applied Mathematics at the University of

Vidbottar's Strand in South Africa where he runs the Robotics Autonomous Intelligence and Learning Rail Laboratory. In 2024, he became the founding director of the machine intelligence and neural discovery mind institute at the University of Vidbottar's Rand which he very kindly said before.

We started. They refer to often there is vits and I'm going to do that too. But welcome Benjamin. Thank you so much. It's great to be here.

Also I'd like to congratulate you because you were named to one of the most, 100 most influential people in AI and time magazine which undoubtedly has led the media to harass you regularly including us. So I apologize for that. But we've been tracking these issues for a long long time.

But we have been interested from the very start possibly because my background is not just in international affairs but also very much rooted in the world of the emerging world. And in how the AI revolution impacts the global south, how it impacts emerging economies, how it is shaped by those.

And what are the things that I've been struck by in reading your work or in listening to interviews with you is that I think you've had a big focus on not just ensuring that Africa is in the game but that African perspectives can shape the game. And maybe that's a good place to start.

Wow. What do you feel that the influence can be or should be? Great. So, you know, as someone who's worked in AI for a while, I've got

“a very clear bias when I say, I think AI is the most important technology”

that we'll ever develop. But be that as it may, when you're developing some sort of key technology, particularly when it's got kind of broad impact and the potential to touch on

many different sectors and ultimately then the lives of many different people,

I think it's very important to think about how it rolls out, how it's developed. And kind of what that, what the different stages of that pipeline look like. So, you know, particularly if we see just how quickly things exploded from, you know, a couple of big moments, particularly when chat GPT did its thing.

It becomes clear that I think different communities around the world, in my case, that's the African community needs to kind of be aware of what's happening and have some sort of role in what's going on there. So, you know, this comes in at many different levels.

This is about, you know, really useful, really powerful technologies.

Are they being used in the correct ways? Are there right tools being used? Are people just following trends or being wise about how they're doing things?

“So, you know, I think just from that perspective, it's important to look at the ways in which”

a technology, particularly when it's coming from another part of the world, which means, you know, at its core, it wasn't necessarily designed with the kind of day-to-day challenges that people on the ground in different areas are thinking about. You know, it's important to make sure that that's done in a sensible way. So, that's kind of the starting point, but I think, you know, this really motivates that, you know,

you need to make sure you've got the right kind of technologists sitting amongst whatever community you're working with,

“but you take this their next level, let's say, are they insights that could come from different parts of the world?”

They could really talk to this as a global mission, as a global kind of endeavor coming from research development

and kind of rolling out into society, ultimately.

And this is really the way that we try and think about particularly AI coming from the African space. You know, there's many open challenges and these range from everything from, you know, the big debates around AI safety to questions around, you know, efficient models and having reasonable amounts of data and the kinds of data centers we have. And, you know, the one kind of, obviously example is, you know, in many places in Africa, we're obviously famously kind of under-resourced in the way we do things,

which means that we've got to think about, you know, either we try and adopt the way that things are done in other places, which, you know, is met with success sometimes, or that forces you to think innovatively and kind of in a frugal type sense.

“I'm a big fan of the idea of frugal innovation and frugal design, a driven by necessity and what could that mean actually for the global discourse around the field?”

Well, there are a million questions that come out of this.

Let me sort of start from the inside out, from within the AI out, both in terms of, structure, linguistics, culture, and then sort of out into the social impact. I do, I note that you've done some work on AI and local languages, and of course, a lot of the functioning structure of AI is linguistics. And linguistics, of course, is not just words, it has embedded within it, culture, it has embedded within it. How different groups approach different kinds of, you know, solving intellectual problems or viewing certain things philosophically.

I was struck by, and frankly, please, to see by that, you know, within your institute, you've got philosophers and others coming together to help shape this. And I'm wondering how, you know, homegrown, how you see homegrown African AI models differing from, you know, US developer, European developer, Chinese developer models to help illustrate this point to broader audience. Right, so, you know, I think, at my core being an academic, I'm, I'm not someone who's going to say, like, oh, we have to throw out the American models or something.

I think, to me, building the right, you know, or the way any, any technology, any field should develop, should be around having diversity and viewpoints. I think that's always been kind of the, the big strength in science. And you, you kind of mentioned this in your question about, you know, that's not to say that, okay, American approach bad, Chinese approach bad doesn't fit, you know, there's many cases where that is the correct thing to do. But, you know, I think very specifically what we're seeing at the moment is this huge amount of excitement around say large language models and, you know, for good reason they're doing stuff that we barely dreamed of in sci-fi.

But, you know, there's, there's various kind of layers to this that we want you to think about. So, is it always the right tool for a problem, we've got this, this hammer that happens to be, I don't know, it's kind of almost nuclear powered and with every bit of extra,

That might actually do a better job, so, you know, this comes in at the level of almost triaging the way to think about problems, but, you know, I think that's, that's at the kind of almost meta level when you, when you dive in like, how should we be thinking about these things?

“I think a lot of that is driven by, you know, the practicality. So, you know, we look at the language example, which I think is the almost poster child example at the stage for kind of differences in African problems.”

We're, you know, you've got something like, you know, well over 2,000 languages spoken on the African continent. Now, if it's English or French or something like that, there's a what we call high resource languages, we would a huge amount of data, and you can play the game that everyone plays, just make your models bigger, throw more data, throw more computer, it's, and like fundamentally things get better. But if you've got a language that's, you know, got a tiny population, or maybe has no written resources even in that language, you've got to think kind of differently about this. And so, you know, let's say you take this problem where you've got a small community of people that speak a language, you want to give them access to translation tools or transcription tools or something like that, because otherwise, how is their bank ever going to.

Communicate with them, right? So it's not even just necessarily about them having access to AI, it's like just a bunch of straightforward digital services might only be available if you've got these kinds of tools.

“And, and, you know, in some cases, if the population that speaks at languages so small they produce so much less data, you could embark on this endeavor to.”

Get like data collectors to go out there and collect a huge amount of data, but that's probably going to take you forever and still won't get you anywhere near the magnitude of what you need.

Or you could take other approaches, you could say, okay, maybe we can find clever ways of augmenting the data we have to artificially increase that data set. Or for the stall, you could maybe say, well, maybe there's other tricks we can employ, maybe we can use different kinds of technologies. And so we're trying to, you know, when looking at problems constrained and slightly different ways.

“I think that's where, you know, these opportunities come from now.”

The advantage is hopefully this feeds back into into other solutions that other people find useful, you know, what one maybe quite different example.

Although there's quite a few I could talk about is, you know, we've got some people in our institute that are looking at different ideas in African philosophy and what they mean about how we should be, you know, thinking about AI risks. And potentially this could have impact globally. We've got other people that are coming at things from more of a lens of psychology and saying, okay, how do we think about morals and norms in a society where there's there's clearly many different sets of these things that.

You know, people come from say different territories or tribal backgrounds or something, they go different priorities in their kind of society, but somehow are like, as humans, we managed to piece these together and and operate in that that setting. How do we make sure that our, we build AI systems that are able to take these into account where, you know, you've got to have some way of ingesting these in some way of treating them as equal, some way of, like, building systems that would like reconcile differences between, you know, you've got to make some sort of judgment, but maybe two different groups of people would have a slightly different way of resolving that.

Now, I think something like that is maybe not the kind of challenges that you'd get coming out of say a mainstream kind of culture or AI lab or something, and so I think that these interesting opportunities which, you know, in some cases might be very different models in other cases might be a different kind of implementation layer that sits between a model developed somewhere else and the application you're trying to plug it into.

But I actually work as the happen on this and, you know, we want to make sure that these problems are thought through before they rise as problems that are actually impacting people on the ground.

This podcast is underwritten and part by the US Embassy of the United Arab Emirates.

It is presented live without editing.

For further information about the UAE's efforts in the areas of artificial intelligence and technology, go to the website of the embassy at www.ue-emBC.org and search for UAE-us-tech cooperation. We thank them for their support. We thank everybody who's supporting this podcast for their support and we look forward to it developing and growing over time because the issue is so important.

“So, you know, I think there is an understanding among a lot of people who've been dealing with AI for a long time.”

But concern that exists even within, you know, the US or European societies that that you can end up with biases built into algorithms, right?

And when you sort of extend it to a broader cultural context, what you can end up with is sort of algorithmic cultural imperialism, right? So, you know, this is somewhere hidden to the end user within this is going to be something it may be a gender bias, it may be a racial bias, it may be something else.

“And there's a lot of talk about that in the north and if, you know, with regard to coming up with models from all parts of the world, I think this that becomes an antidote to that.”

But, you know, it also strikes me that that's a kind of a narrow way to look at it and that every different part of the world has something new to bring to it.

And so, you know, there may be a kind of, I don't know, you know, African secret sauce that's being brought to AI developed there, because people think in a different way they're exposed in different things. I've seen you even talk about, you know, African biodiversity and the models that it gives people to work with as a potential driver of this kind of thing. And by the way, none of this is to say that institutes like yours may not be simply doing cutting edge work and the pure science of it. So I'm not suggesting for a moment that the work being done in Africa is all about being uniquely African, but, but I'm just wondering, in terms of, you know, the special upside.

“I'm ensuring that it's a global conversation, what, what, what are you saying?”

Well, you know, I think the, the kind of, I find it very interesting looking at, you know, what are the kind of inherent, I don't even know, you call it pros and cons, but just differences between, you know, the way things are done in different places. And like, even in terms of culture and what this almost means for, for like a kind of research community, one of the initiatives I've been involved in was a movement that a group of us started back in 2017 called the Deep Learning in DARBA, which was,

it's now become the biggest machine learning summer school in the world and there's, we have this huge annual event and there's something like 47 African countries that her satellite events throughout the year as well. But a big kind of value that we often talk about coming from South Africa is this notion of going to, which is this kind of often described as a spread of community or the idea that I am because we are. And, and many people almost look at it as a bit of a cliché thing, but, you know, that's just almost subconsciously gotten so baked into this movement that we'd put together that, you know, there's this big drive from the community, everyone's feeding back into it.

Everyone's trying to help others and kind of spread the word and grow the community and help people in their own local communities, so much so that over the years we've been asked to help set up kind of similar networks in South America and Eastern Europe and Southeast Asia. And, you know, that's something quite different from a technical research idea that's almost like a kind of way of being or a philosophy just for groups of people coming in together collaborating. And, and I think this is, this is also something interesting, like that, you know, you just find something that, again, nothing to do with the tech world, but when you import certain ideas just about how different societies and groups organize themselves.

There's things there that can give, you know, whether it's competitive advantages or just changing the way that things might be done in other places for the better.

Yeah, well, we're also seeing different models evolve, you know, and I think, you know, we saw with the internet, you know, there was a European model where based on their view of community and society, you know, people had to opt in to using services that there's an American model, which was more corporate driven, where people had to opt out.

There's Chinese model where everything was owned by the state and that you would do what the state wanted you to do.

And now we're seeing the same thing in safety rules and and other kinds of rules applied to AI and even within countries, you know, we're having to bait right now in the US because the Department of Defense has said, we only want to AI that will let us do anything. You know, including surveil people and have autonomous weapons by forms and so forth. And I'm wondering wondering what there is, you know, in these conversations you're having in Africa that suggests maybe a different model or differences with those models.

“So I think one thing that is, you know, first of all, again, that just the principle of having different models, the more in a sense, the better, you know, in the giant sandbox that is our development.”

But, you know, one thing that we've seen is that a lot of kind of what's happening in the AI space that's really been a story of the last not even ten years or so that it's gone from very little to a very connected thriving ecosystem has been about grassroots movements.

And I think a lot of that does come from, you know, just the differences in kind of infrastructure and the relationship and the capabilities of governments across the continent.

But you've really seen this kind of almost awakening of people at all different levels, whether they're young students or academics or in startups or big corporates and so on. Just forming these very loose networks largely organized by various kind of social media type platforms that they're then subsequently using to produce novel research that's winning awards and produce kind of crazy ideas that are attracting a lot of funding or new startups and so on.

“And I think even just just that that there's notion of kind of movements that span entities that are kind of borderless within the continent and in fact beyond.”

I think that's something of a different model to what I've seen before that I think is quite exciting that, you know, we can attribute a lot of the big kind of successes that we've been seeing over the last few years to just kind of coordinated networks of people in very different circumstances, very different countries. Many just doing these as kind of side-side side hustles. Just producing interesting new ideas that are kind of building in momentum so that this whole, for example, movement around African natural language processing really came out of a network called.

So it's a, which is a completely self-organized network like this, but it's now received quite a lot of funding and a lot of awards and so on, which again was a network that came out of our deep learning and dober movement that we we set up, which was another grassroots movement and so seeing these kind of compounding effects and drawing in people from around the globe that's again done in built in a very different way to being centered in big research labs or centered around single universities that that are either collaborating or competing.

“I think that's an interesting different model of innovation that's now kind of the main engine room or distributed engine room or the continent.”

You have a lot of stress, you have the feeling of stress, you have the feeling of stress and then you have the feeling of stress. You have the feeling of stress and then you have the feeling of stress and then you have the feeling of stress.

So we are going to see a new one in the world. I would like to say that I would like to finally get into two questions about impact applied AI impact within the societies.

I was at a meeting a couple years ago, it was like a World Bank IMF meeting here in Washington and I was listening to some guy and he was going, well, you know, all those call centers in India.

You know, they're all going to move here to Africa and so these are this is going to be a big job creator here in Africa.

“And I think it was related to the sense that there was the infrastructure to handle this at the time.”

And I was there with a couple of guys who were more into where AI was going at the time and they said, no, they're not. Those jobs are all going to go and be done by machines. And so, you know, on the one hand, I think, you know, there is this sense that AI will produce labor market dislocations. And that some of the people who will be affected by that are, I don't know, aspiring lawyers who want you once used to spend the first seven years of their career doing research that now machine can do.

But a lot of it has impact on a broader worker basis, particularly in developing societies.

If you go to any university here in the United States right now, and I find this bizarre, since that jet GPT moment you talked about was couple two and a half years ago. AI skepticism is off the charts. I mean, there's almost a generational thing where it's like, Gen Z is like, AI bad. It's not, you know, there's no, you know, there's no nuance to the argument. And I'm just wondering what the sense in the communities that you're dealing with is of AI as a negative disruptor and of AI skepticism.

“So I think it's interesting. We've got generally, you know, the full mixed bag, I think is everywhere.”

I think there's generally a lot more optimism around here, you know, when we, there's clearly a lot of, you know, famously a lot of problems in Africa.

And, you know, I think there's quite clear views on that AI can be used to address many of these challenges. When we, when we look at kind of the job market effects, they, they are a lot of fears. I think in many ways we're in a, in a slightly better position because, you know, some of these countries that take my own country South Africa, for example, that's something like a, like officially, that's 31% unemployment rate. And which means we're very sensitive to discussions around employment and kind of protecting jobs and this kind of thing.

“And I think it's, it's something that people are more thoughtful of it. It's a lot harder to fire people.”

It's, you know, even like I'll be corporate investor lot in kind of training pipelines and that sort of thing. But there are certainly fears. I mean, you know, you can look just at, for example, in the education space. The number of people that, you know, are saying, okay, if students are using this, no, we've just got to switch it off. We can't let them use it because then they're going to cheat on their assessments, you know. And I think there's, there's particularly outside of the, the direct text here.

There's a lot of people that, you know, particularly seeing it as a, I think a lot of the challenges come from like an affront to kind of people's identity. And like they're positioning in society. But I've seen less of this here than I have in other parts of the world. I'm not sure why that is. It might be a slower adoption that might be, you know, there's quite a lot of people are doing a lot of work around like demystifying the technologies and trying to get people to embrace it. But it might also come down to like a culture of, you know, we go through many generations of new technology pops up, you know, we got now social media.

We've got, you know, there's WhatsApp, like, do any of these things give us abilities to reach out to rural communities and help them with things. And I think we've maybe got more of a track record in people trying kind of a social innovation type things around new technologies to say, how do we help people that that perhaps there's more of just a trend of like, you know, some of these things work, some of these things don't work, but it's good people doing it at the end of the day. And I suspect, you know, some combination these factors leads to, and maybe I'm just getting a kind of wrong view on this, but what I generally perceive as a, I don't know if it's a more positive lens, but certainly there's less negativity in that sense.

Well, my, my final question was going to sort of turn us in that more positive direction. One of the things that I've seen dealing with development issues for a long time is the role technology has played in.

It's rural areas and all of a sudden you've got drones evaluating fields and dropping, you know, fertilizers or seeding or doing other kinds of things, or, you know, another phenomenon, which I think has been kind of great is, is cutting out the middleman for people who do, who create things and, and they've got the ability to go and market directly on the internet by themselves or by a.

Collectives on a global basis and I guess one of the, you know, I mean, we're seeing AI, I talked to and Israeli.

“Entrepreneur investor, very big investor and in companies and one of the things I said, what are you looking for?”

They said, well, I'm looking for the three person unicorn and I was like, well, what's a three person unicorn and he said, well, it's a company. That's using AI to give three people the ability that once, you know, 50 people had and so it can grow very rapidly with very few people.

And we're seeing more of that and and so, you know, to me that that creates opportunities also in emerging economies and so I'm wondering, where where you anticipate the leap frog.

I think that's a great question and it's, you know, this is the beautiful thing with AI is like just what it can touch on and you use back about many of these things. I mean, the kind of everything from rural health care to improve service delivery to kind of streamlining bureaucratic processes to updating kind of educational systems that are really struggling with scale and that kind of thing.

“Like, like all of the financial payment payment systems where it might go directly into a bank account that lives in somebody's phone, right?”

Absolutely. So we've, we've got pretty high self imprinted penetration across Africa and I think all of these are a big positive effects.

Possibly one of the biggest things is going to be that there's a lot of initiatives now that are signing up with people say, are we going to train this many hundred thousand people or this many million people in AI and I think, you know, particularly with kind of unemployment with small economies and this kind of thing. I think there's a lot of that to me that's where there's a big leap frog potential rather than just kind of a gradual improvement of things is that if you can. Equip a bunch of people with just a set of tools that have very little friction in them and give them their ability to create and create new opportunities for themselves.

This is something can be done kind of at scale and, you know, then you've got this kind of that downstream effect of those people in innovating.

“I think that kind of approach is probably where we're going to see the biggest jumps because that's, you know, maybe the hardest to predict and that's the easiest to scale in a sense.”

Well, it's, it's, it's extremely interesting and it is an area that does not get enough discussion and clearly you're a treblezer in this area and. I'm delighted that we've had the chance to talk to you even briefly on this. I feel conversations have been superficial and I apologize for that, but I also hope you will see it as a first conversation that we can come back to you and continue it. We'll continue the discussion in the months and years ahead, but for now. Thank you very much. Thank you. Thank you very much. And, and thank you to everybody for listening. We're here every week doing.

Conversations like this on sale consciousness also on our AI energy and climate podcast which I encourage you to listen to. And all the other podcasts we do here at the DSR network. But for now, thank you Benjamin. Thank you everybody. Bye bye. This was Siliconjustness, a production of the DSR network.