263 – Earthkeepers versus AI Empires (Part 1)

[00:00:07] David: Hi and welcome to the IDEMS Podcast. I’m David Stern, a founding director of IDEMS, and I’m here with Kate Fleming.

Looking forward to this discussion, I’ve just come back from an event and am so keen to discuss.

[00:00:20] Kate: Hi, David. Do you want me to set it up and then I’ll ask you questions about it or would you like to introduce it?

[00:00:26] David: Go ahead.

[00:00:27] Kate: Okay. So basically you attended this Earthkeepers versus AI Empires event in Zambia, which was intended to be part of the RightsCon event, which was cancelled for reasons we won’t get into, we could do another whole episode about that. But very specifically, this was what was intended to be a sidebar event to this main event, but ended up being your big focus.

But you went to this event in Zambia and it really brought into focus for you that there is an emerging community that is quite knowledgeable and educated around issues surrounding AI in ways that you hadn’t understood, or really I don’t think either of us had really seen was out there in the way that it is.

And then the other piece of that, which I’ll just lay as like a foundational element of, I think, what will shape a lot of our conversation, is that much of this has been shaped by the book “Empire of AI” by Karen Hao, which I think is now out in enough places and has been read by enough people that I don’t think this will be unfamiliar to people even if they haven’t read it.

The way she tells the story, she is such a good storyteller and able to articulate things that have often been quite complex or felt very diffuse in ways where you see the connection, you see the narrative, you see the forces that have shaped a lot of where we are. And that none of this is inevitable.

And so, I guess, just as a starting point, what is the background of this group? Where did it emerge from? You know, I see that there’s a website, there’s obviously some history there, but just to give us a bit of context, could you just talk about the group a little bit?

[00:02:17] David: Yeah. So, I was at this event in Ethiopia at the end of last year, this was a Pan-African event. And at that event, a lot of the people who actually created this event were there and were in the room, and I was part of that. And I was part of it without feeling I knew what was happening, because a lot of this was activist groups who were saying what was happening. And what was relevant in this African focused event, which led to this Pan-African declaration, was relevant more broadly. And so they wanted to get these discussions happening on the boundaries of RightsCon.

And a guy called Jim Thomas was one of the people who took the lead. And other people, including members of the ETC group and broader members who are involved in, well, human rights related to digital technologies more broadly were interested in how this relates to agroecology and the issues around food sovereignty.

And so that’s really where this event was born. And I was sitting at the table as this was being discussed and it emerged. And then it took shape in a way that I would never have expected, mainly ’cause I’m not in these spaces, I’ve never been in activist spaces before. But there are a number of people who are used to recognising when human rights issues are, you know, when we are getting to the boundaries where it is really unacceptable and who then know how to act on this.

And they then really recognised the issues around, I suppose with the Zambia focus, these big data centres. Zambia as one of the world’s largest copper producers is now becoming a natural space which can be the home for these big data centres. And they actually discussed how to educate people on the fact that not only this is really rather dangerous for the environment and on the actual human rights around some of this, but that it’s also unnecessary. And that was part of what came out very clearly.

So the fact that it was in Zambia was not by chance, it was in Zambia because of these big data centres.

[00:04:52] Kate: So I guess what I’m hearing in what you’re saying is that a lot of the origin point of awareness really was driven by data centres, which we’re seeing everywhere globally, where I think this is happening in the US, it’s happening in the UK. You see that suddenly there are these voracious entities even just competing with a local community for water, for power, for whatever the things are that you need to live. 

So on a quite immediate local level – which I think is very often how politics happens, I think all politics is local, it’s how you get people activated – it’s interesting to hear that that’s a real impingement on people’s quality of life, on the cost of resources, on resources that can be quite scarce, like water. Suddenly water, and clean water, is in demand in ways that are just not able to be borne by a community.

[00:05:52] David: Well, and this is where Zambia is such an interesting case because they have a relatively large proportion of the world’s clean water, similar to copper. They are actually very resource rich in the things which are needed, which is why they have become this magnet, if you want, for the technology at its biggest scale.

[00:06:16] Kate: And it’s an interesting observation because I think we often avoid terms that feel heated and politicised because we are builders, not activists, but this is, if we’re gonna identify colonialism and what colonialism is, you can’t see anything here other than that. This is ground we have covered before, we know what this looks like when outside forces come in.

It might not be countries, but a company but basically the scale of a company is equivalent to a colonial force coming into countries. That’s how people are experiencing it, where they are seeing their natural resources, their quality of life, these various things just being syphoned off for somebody else’s good with little to no benefit experienced locally.

[00:07:04] David: And this is the key point. I’m not normally in these spaces, but this particular case is a really interesting instance of where development and economic development is being used in a way, which is, potentially, damaging a country at a scale which is enormous. Zambia, the scale of the mining, the scale of the data centres, the scale of what’s being built there is on an order of magnitude, which the continent hasn’t seen before. And therefore, the implications of this are really resonating out across the continent.

[00:07:43] Kate: So this group that convened, is this primarily an African convening? Is it a global convening?

[00:07:51] David: It was global. It had people from all over the world. It had of course a large number from Africa, which, for the convening being hosted in Zambia, was absolutely appropriate. It had a large number from Southern America, Brazil particularly was very highly represented. Of course, North America was highly represented, Europe was reasonably represented as well. Asia only had small representation, but it did have a few people there and were represented.

So, it was conceived and constructed as a global convening. And it was, as I say, for me, a space that I’ve never been in before, which was led by activists and by people who were well versed in fighting for their rights.

[00:08:42] Kate: Okay. Because Karen Hao’s book is going to be a theme through this, there’s a passage that I think is really interesting that might be a helpful reference point for us. And I think part of how I would wanna frame it is that the rapid change of pace of technology is so profound, but at the same time it means actually that people have been upskilled in like anticipating harm. So, because we had social media, because we have these things, we’re already dealing with the consequences.

I think when it comes to AI in particular, people are on high alert where it’s like, okay, we saw these harms that came before and we weren’t ready for them, and they kind of got embedded in ways that we feel like it’s very hard to roll back. And I think part of what we’re seeing here is a rise of the equal and opposite pushback force much more quickly than we have seen in past movements of change. Whether that’s industrialization, whether that’s the early commercialisation of the internet, whatever that looks like. So I just wanted to read this passage because I think that it feels foundational to what you experienced, and I’m curious to hear how it lands for you.

So this is from Karen Hao’s book: “In their book, Power and Progress, MIT economists and Nobel Laureates, Daron Acemoglu and Simon Johnson argue that every technology revolution must begin with a rallying ambition.” So I will just stop there for a moment. I think this is what we are experiencing, it is “oh, AI will change the world, it’s going to do all these amazing things”. There’s this grandiose talk that is incredibly inspiring and is exciting. Who wouldn’t wanna be a part of that?

And so its promise is of a technology benefiting everyone that puts in motion the long journey of amassing enough talent and resources to turn it into reality. But the authors conclude that technologies are not inevitable. “The ability to advance them is driven by a collective belief that they’re worth advancing. The irony is that for this very reason, new technologies rarely default to bringing widespread prosperity” – the authors continue – “Those who successfully rally for a technology’s creation are those who have the power and resources to do the rallying. As they turn their ideas into reality, the vision they impose of what the technology is and whom it can benefit is thus a vision of a narrow elite imbued with all their blind spots and self-serving philosophies. Only through cataclysmic shifts in society or powerful organised resistance can a technology transform from enriching the few to lifting the many.”

And so I highlight that because that feels like what you are stepping into. When you talked about this event and you were so energised, you already see people thinking about how we are really going to do this.

[00:11:34] David: Absolutely. And it’s very interesting that the resistance is what a lot of these people in this event were talking about. And I’ve never been in events where that is the focus before. But it was actually really heartening in a way that became clear as one of the things that came out was that big resistance actually was galvanised by these data centres. The data centres are the physical system, you know, one of the terms that came out and was used quite a lot is there is no cloud, there’s only the earth.

The data centres are the physical embodiment of the cloud, for these systems to run, they need, they’re data hungry, they’re water hungry in ways which are now appearing and being constructed in this way, which is competitive.

While I was on my trip, Kenya announced it was not going to proceed with the data centres because it didn’t have the power infrastructure. And Kenya is known within Africa as being at the forefront of technology. So for that announcement to come is a state of just how ravenous these data centres are.

They had the plans, but once it actually was worked out that this meant that the Kenyan grid would be basically totally overloaded and unable to sustain this without huge other power investments, the Kenyan government, which is very much behind this, stepped back. The data centres and the amount of energy and water that they are requiring is at an order of magnitude which is actually galvanising this resistance and bringing people together in a way that I wasn’t really aware of.

That to me was very new, but it was really interesting. I didn’t go into that meeting with an understanding of how that is a rallying point, but it is really clearly emerging as the physical impact of this narrative that we need more and more data and compute.

[00:13:56] Kate: So that brings me to a question, we’ll see if I can frame it well. I think there’s one layer of activism which is “I’m just reacting against what is”, and it’s often like sort of the Luddites, and we’ve come across Luddites before. They’ve kind of been reframed by history as these anti-progress people, when actually they were just really saying, well, how do we do progress in a way that doesn’t hurt people, that’s inclusive, that’s bringing everyone along? So that’s reexamining what a Luddite is.

But there is one model here where it’s just that you have people reacting to data centres. There’s another model of activism where it really means that people are starting to interrogate the whole system, where they’re saying, well, wait a second, why these data systems, why is so much processing power needed, why do we need all this energy, why all of these different problems? So I think you found yourself in an interesting position, and I guess this is part of my question: where did you feel people were already quite aware of how AI is problematic and the data centres are just the manifestation of this whole system that has a lot of things that are problematic? And where did you find that people hadn’t connected the dots in ways that you were able to? Were you able to see pieces that advanced the conversation or shifted the conversation?

[00:15:16] David: Well, it was really interesting because there were of course people in the room who were proud of not using AI, and there were others who were AI users who were saying, well, actually, I’m able to use it in ways which are more effective, which are improving what I’m able to do and what I’m able to contribute to society. And there was a whole range in between.

And one of the things which, again, right from the beginning I was able to sort of help articulate, is this division between AI as a technology which has been going for 80 years and where there have been a whole sequence of improvements over time and it’s a tool and it is not good or bad, and implementations of AI, as explained in the book, that are extractive, making choices which are unaligned with a lot of the communities that were represented and which therefore they were correct in lining up as resistance to them.

And I had wonderful examples talking to, again, a context I knew nothing about, indigenous American communities whose languages had been studied to protect them, but actually not in a way that was respectful or actually aligned with them, and where that power and ownership was being taken away from them.

And so it was being brought in and technologies were being used, artificial intelligence technologies were being used as a way to take ownership and control of what was, of course, really important culturally and theirs. So there were really interesting discussions that emerged from that, about an example which I often use as one of the potential benefits: if we do AI right, it could protect the diversity of languages across the world in ways that would be useful.

But they were saying that they had experienced this not as protection, but as capture. And that’s really interesting how foundational that was. I know bits of work that are possible on this, but I hadn’t been that involved or aware of work that actually happened to try and use AI to understand some of these native languages. It was really interesting to see how that machinery is already being used in ways where it is capturing.

That was really new to me and highlighted that there’s this whole spectrum of people, from users who naturally were very happy that I was there, to the people who were very anti, who also appreciated the perspectives that I was able to bring on saying, well, this is the tool and that’s the empire, that division of what is the tool and what is how people have chosen to use it.

One of the bits that was so interesting was the role of academia in the empire, that actually a lot of academia has been, if you want, captured by the empire. And that’s something which is in the book as well, and I articulated in a way that I’d never been able to explain as well as Karen did, the way she was then able to bring this out.

And what was so powerful was that I was having a lot of these discussions over the first few days, and sort of having a position that I was trying to explain on this. And then Karen presented actually on the third day and she articulated everything I was trying to say and more, much better. Her communication skills on this and the depth of her understanding are just so powerful in actually telling this story and explaining how deep the problem is.

[00:19:23] Kate: Yeah, and I think, to come back to that quote that I read before, there is an element of “it is neither all good nor all bad”, although bad is certainly an ascendance along with good in ways that are hugely problematic. But I think AI has been galvanised by a really ambitious vision and getting people excited about it and this promise of solving these huge problems.

Those are things I think we can all get on board with and I like to envision that future too. I think what you are helping to sort of distil down into component parts is, well, what is actually contributing to making that happen legitimately, and what is not, and where is it smoke and mirrors, and where are there a lot of interests that are co-opting this quite positive story to actually possibly diverge into some path, where they’re forking very far away from ever actually getting to that more meaningful thing, even though they’re continuing to own that grand vision that makes it like, well, “”if you stand in the way of data centres, you stand in the way of progress”.

And I think that’s where you were able to sit, as a technologist, as a data scientist, in all of these spaces and say, okay, “fact, fiction, fact, fiction”.

[00:20:42] David: That’s right. That these data centres are not needed for the AI that actually you are wanting to use is one of the things I was able to really explain in particular cases to people who were really confused by this and who say, “you know, I use AI, it’s really useful”. And I say, “okay, well what do you use it for?” “Okay. That doesn’t need this data centre.”. This is what’s so exciting about the technology. The advances which have happened, if we actually focus the effort in a different direction, everything that people care about could be achieved without the big data centres. It’s not to say without any data centre, but without this enormous need for compute and data in the way that it has been articulated and in the way that the narrative has driven.

[00:21:29] Kate: Yes. And I will say, I think we’re going to save that for a later episode because there’s a lot to unpack in that. But just to stay focused on this event that you attended.

[00:21:39] David: We are out of time, we haven’t really dug into the event as it unfolded. Should we split this up and actually have a second episode where we dig into actually some of the deeper things that came out of the event and what unfolded?

[00:21:55] Kate: Yes, let’s do that.

[00:21:57] David: Great, okay, speak again soon.

[00:22:00] Kate: Thanks, David.

[00:22:01] David: Thanks.