How Cloud Giants Wield Their Power Against Us All w/ Cecilia Rikap

As Big Tech’s power continues to grow, it's more important than ever to understand how companies use their control over cloud infrastructure and services to shape more than just the digital world. Cecilia Rikap joins Paris Marx to discuss how these networks of corporate power operate, including how they are systematically used to influence global politics, the economy, and how society views the world.

Cecilia Rikap is an Associate Professor at University College London’s Institute for Innovation and Public Purpose and the author of The Rulers: Corporate Power in the Age of AI and the Cloud.

The podcast is made in partnership with The Nation. Production is by Kyla Hewson. Support the show on Patreon.

Also mentioned in this episode:

• You can now pre-order Paris’s new book Hyperscale.
• Cecilia and Paris worked on a white paper offering a roadmap to reclaiming digital sovereignty.

Transcript

SPEAKER_00 0:00

It is also that they have this unique panoptican power of the global economy that enables them to do multiple things. But ultimately, the most important thing is that enables them to control, to plan in an authoritarian way the trajectory of that of our economy and our lives, therefore.

SPEAKER_01 0:40

Cecilia is an associate professor at University College London's Institute for Innovation and Public Purpose. She's also the author of The Rulers: Corporate Power in the Age of AI and the Cloud. Today we have massive companies like Amazon, Microsoft, and Google that are not just huge tech companies in their own right, but also control the infrastructures that so many of the services we rely on, so much of our data are stored on. You know, if you think about it, it's not just our personal files that are on the cloud and in all the data centers that make up the cloud, but also what so many corporations and governments do as well. You know, it's all on those servers, it's all on those services that, you know, we collectively term the cloud. But because of that, companies like Amazon, Microsoft, and Google gain a lot of power over not just us as individuals, but the wider society because so much is hosted on their servers, so much relies on the service that they provide. And that not only allows them to get insights into what so many other companies and even governments are doing based on how they use their services, but it also gives them leverage too, leverage that they can use to get things that they desire and to even shape what so many companies and governments do, how they use technology, even how they think about operating. It's a huge amount of power that we don't often consider. And that is why I wanted to have Cecilia back on the podcast to talk about her new book, where she really digs into the deeper aspects of this, right? Because it's not just looking at the cloud itself and what these major companies are doing with it, but also how so many other companies that we don't consider tech companies are relying on this cloud, you know, on this service as well, and how that shapes the way that they act and the way that Amazon, Microsoft, and Google and the actions that they take can then shape what so many other companies throughout the economy actually do and how they operate. So I think that this is a really important conversation to actually understand the scale of what is going on here. Because if we ever want to challenge the power that these companies have, we need to understand actually how it works and the degree to which their tentacles have gone throughout society, throughout the economy, and the many ways that's going to need to be reined in if we actually want to disempower them to take back power for ourselves. So I think you're really going to like this conversation. If you do, make sure to leave a five-star view on your podcast platform of choice. You can share the show on social media or with any friends or colleagues who you think would learn from it. And if you do want to support the work that goes into making Tech Won't Save Us every single week, so I can keep having these critical, in-depth conversations that explore so many different aspects of the tech industry and the way that it affects our lives. You can join supporters like Mike in Stockholm and Marcus in Houston by going to patreon.com slash tech won't save us and becoming a supporter as well. Thanks so much and enjoy this week's conversation. Cecilia, welcome back to Tech Won't Save Us.

SPEAKER_00 3:31

Hi, how are you, Paris?

SPEAKER_01 3:33

Yeah, I'm well, thanks. You know, happy to be talking to you again. So that always brings me some joy. Um of course, you have this fantastic new book out called The Rulers, you know, where you dig into the way that these massive tech companies really gain a lot of power from, you know, the cloud services and the AI services that are increasingly ubiquitous, that so many companies and people rely on. And so I just wanted to start with that kind of broader question, right? You know, you've really focused in on Amazon and Microsoft and Google in the book. What is unique about how those companies exert power over the economy and society today?

SPEAKER_00 4:11

Well, that's that's an extraordinary question because actually, when people speak about AI these days and the cloud and so on, they it it's like you feel that it's all over the place. You have the open AI, the anthropic, and then deep sea comes, and you have Chinese companies and Mistral, and they're all these AI companies, and they seem to be kind of the ones that will displace the big tech companies. And when we say big tech also, that has become a bit of a buzzword. And what do we really put inside that label? So originally, like if you were in France, you would have said, okay, GAFAN, so Google, Amazon, Facebook, then Meta, Amazon, and Apple and Microsoft. But then you also have other companies that should be included. And what is the difference really? Is it that we can label just the companies according to and say the big tech are the ones that have the largest market capitalization, for instance? And then now we have to also include SpaceX. But then we look into the financials and we see SpaceX not making money, actually losing money. And the same happens with the AI companies that still haven't conducted their own IPO, so they are not public companies yet, but still we do know that they're struggling and they need to find a business model eventually. And I would say, I would still put that into a question mark, and I will come back to that later on, because the book is precisely about understanding that behind all this mess of companies, there are some that are always profiting. There are some that are actually calling the shots and deciding what type of technology we have. And those are not necessarily the ones developing the latest iteration of an AI model, but those that control the whole space, the whole ecosystem where the AI models, but also all the other digital technologies, are more and more not only being connected to each other and exchanged, but also produced, consumed. And this is the cloud. So basically, I've been studying digital technologies for over a decade now. And the first thing that I came, I realized was that these companies, Amazon, Microsoft, Google, were very early on invested in AI. And it was like everything about AI. At that time, they and we and everyone was talking about AI like machine learning, deep neural networks, and so on. Those were the latest approaches. But on top of that, work on AI that is still due, and Gemini from Google is also among the leading AI models, for instance. The three companies eventually moved towards focusing more and more on the cloud, not only as infrastructure as a service. And of course, infrastructure matters. And part of the story that we've seen lately and we see these days with data centers, and you are focusing a lot on that, and it's critical. It's critical because infrastructure, of course, matters. But it's not just about the cloud as infrastructure as a service. It's not just the cloud as where we are storing data, where the models are trained, or where the models are run, but it's about the cloud in understood as the platform of platforms, if you want. The place where every digital technology is, again, produced, exchanged, and consumed. I always say this. I'm an economist by training, and in economics, you get this like Econ 101. We have like in capitalism, there are three separate spaces: production, exchange, distribution, the market, and then consumption. Each of these economic spheres are supposed to operate independently and are supposed, therefore, to be operating under their own rules. The logics that operate behind them, inside them, will be determined by each of the spheres independently. And apparently, in the case of the market, it will just be the autonomous uh process that brings together all the actors that comes up with resulting outcomes. All this is a complete fiction that doesn't describe how the cloud works and how capitalism more broadly works, actually. And the cloud is very illustrative in this respect because, again, while we are, for instance, uh, if I'm a startup and I'm developing a service, that service, in order to be developed, will be consuming software as a service, data as a service. So at different types of digital technologies that are not just processing power, and all of this will be connected like Legos. And the final piece will be my application. But in order to sell that application, in order to get others to deploy, use that application, I will also use the cloud. So even Meta, that has its own AI models, then offers the models as a service on Amazon, Microsoft, and Google Cloud, the same with DeepSeek, uh XAI, so before, like now, SpaceX also started developing these agreements with cloud companies. Because sooner or later, if you want to be part of the ecosystem, you need to be on the cloud. So this is why the cloud is so central. And this is if you want like one part. The other part is what happens with all the other companies, organizations that are not originally from the tech sector. When they start adopting digital technologies, and AI has become a sort of like very good marketing strategy for that, because if they were not yet on the cloud, now they even feel a larger pressure to be there. Why? Because there is where they will be able to use AI models as a service, creating an application or just prompting the models. So sooner or later, organizations from the most diverse sectors, industries, governments, again, like schools, hospitals, they migrate to the cloud of Amazon, Microsoft or Google. And this not only gives the companies and also, therefore, the US government to some extent a kill switch that determines whether the hospital can run or not, uh, whether you can get an x-ray, whether they will know that you are waiting for an appointment or not, because everything that we do these days relies on digital technologies. And if these technologies are used as a service on the cloud, it means that everything that we are doing is directly or indirectly connected to the cloud, like our conversation as well. But it's not just about the kill switch. It's not just about the what if one day they cut the service. It's also about the ruling, the decision-making power that then Amazon, Microsoft, and Google get over our societies through these technologies by deciding how we govern every single space of our lives. And this is why it matters so much.

SPEAKER_01 10:48

I think that's such a good like encapsulation of the argument that you make in the book, right? And like looking at the different elements of it. Yeah, I should really look and see what uh cloud service the recording software I'm using for this is on. Uh so I can have a better idea, right? They're all going to be on some cloud service, as you say. I don't think I'm gonna be able to fully avoid that. But I would not be surprised if it's Amazon or Google or something, most likely Amazon, because you know, it's the dominant one of the dominant players. Um, you know, before we we dig deeper into what you're saying there, can we talk a little bit about what the cloud is like in in practice? You know, because you talk about it not just as a platform, but as a marketplace where even a company like Meta offers its products, even you know, even though it's it's trying to sell them separately, but it knows that it needs to be there as well. So, like if people are thinking about the cloud, it's easy just to imagine, you know, some fluffy things in the sky. But there's a really material and kind of digital piece to it as well. Like, what is the cloud really?

SPEAKER_00 11:45

Well, absolutely. And the first, the first link that we should do there is of course, the cloud relies on the data centers and therefore on electricity, the consumption of electricity, drinking water, a lot of resources. So the cloud, although it seems to be like, again, just on the up on the sky, it's deeply dependent on nature, it's deeply dependent on extraction from the earth, and it's also deeply dependent on extraction from us as humans in our workplaces, in our everyday life, because the cloud exists and the history of this modern cloud, actually. And I say modern because also in the book I go back to like old examples from General Electric and how it was also offering the possibility to share infrastructure and why other companies at that time already were uh using this opportunity to only use pieces of infrastructure. But today, modern cloud relies, and if you go back to the history, it relies on Amazon.com, the original business from Amazon. Amazon was growing so much, they needed more and more data centers. But it was not just about the data centers for all the data that they were capturing from all of us through the marketplace. It was also about all the models, all the AI models that they were using to run this platform for identifying like the recommend the recommendations that is done with machine learning, setting prices of every single of the million of products that are sold on Amazon.com that uses AI machine learning and all this required also a bunch of other complementary services. One, as a user, we just connect and we just, I don't know, prompt ChatGPT or Cloud. But behind that, behind that interface, there is not only an AI model, but a lot of different pieces of software and layers of hardware interconnected. And all that needs to be interconnected somewhere. And the virtual space where all this happens, like the oven of digital technologies, is the cloud. And the cloud, as I was saying, I usually use the supermarket as a starting point to compare with something that it's like everyday life for everyone. So if we think of the supermarket, and I live here in the UK, so I have a Science Burris Tesco nearby. So I go to Tesco and I buy whatever I want to eat that evening, let's say spaghetti. Those spaghetti would have been produced somewhere else. It could actually be in Italy if they are Italian ones at a factory of an Italian company, and then they were shipped to Tesco or Saintsburg. And then there is this separate moment, the market, the supermarket, where I go and buy the spaghetti, I buy the tomato sauce, I buy the cheese, and then I come here to my place and I prepare my meal and eat it. And I'm so descriptive because I want to explain how all this actually, when we move to the digital world and digital technologies, and here we are speaking from a cybersecurity software to a facial recognition software to an ERP and enterprise resource planning. So the systems, the uh the software that companies use to plan their value chains, to control workers, everything, everything is produced entirely inside the cloud. When I say entirely, this means several things. One, it means that the developers of all these companies, so me as a cook preparing my meal, I will be doing it inside the cloud. It also means that I'm buying the spaghetti inside the cloud. It also means that those producing the spaghetti are producing it inside the cloud. And when you produce your spaghetti inside the cloud, you do it with services. So let's say with flour that you don't know where it came from, you just purchase it as a black box, you use it, and it's indispensable for your spaghetti. So every time you're producing espaghetti, you're purchasing that flour. But unlike what happens with uh the spaghetti factory, that basically, if it has a problem with one supplier, it can go and purchase flour for someone else. Once you build all your architecture on the cloud, it's not automatic. Sometimes you cannot even find alternative services because everything is produced. I for this book, I conducted over 100 interviews. And when I was in and from and I interviewed people from the tech sector, but working not only for big tech, but also for uh other big companies. And someone from Amazon Web Services or from Amazon's cloud told me my job is to create a stickiness, which basically, from a competition policy perspective, is this person's job is to create a lock-in, to contribute to creating these barriers so that it becomes almost impossible, even if there are three companies, to move between the clouds easily. And you can do it. In principle, it's not, I mean, it's not that it's forbidden, but your architecture ends up being so intricately so so entangled with all the services that you purchase without truly accessing the code, that the work that it would take you, the money that it would take you as an organization, just makes it impossible to leave. So this is why also it's a unique way of producing, but it's also a unique way of producing because there is something else. If we have everyone producing their digital services and consuming digital services on the cloud, this means, I mean, and and I don't want to enter here into the philosophical discussion, but just to continue with an easy analogy, this means that the global economy has a digital twin that is operating on the cloud, right? And this digital twin is giving Amazon, Microsoft, and Google an extraordinary oversight over not only the digital twin, but the whole global economy. And this is also unique of the cloud, and this is also unique of how digital technologies are being produced today, and results in extraordinary power. It's not just a choke point, it's not just a bottleneck, it's not just that if you want to produce digital technologies sooner or later you need to go to the cloud, unless you are in China and you have your own Chinese versions of this same story. It is also that they have this unique panoptic and power of the global economy that enables them to do multiple things. But ultimately, the most important thing is that enables them to control, to plan in an authoritarian way the trajectory of that of our economy and our lives, therefore.

SPEAKER_01 18:10

Yeah, I think that very clearly shows what we are up against in trying to figure out what these companies are doing and really taking on their power, right? And you explain really well how they gain so much power from the fact that so much is happening on their clouds. And that's where I wanted to turn next, right? Because it's not just Amazon, Microsoft, and Google who are running things on their clouds. It is all of these other companies who are putting their businesses, putting their operations within the clouds, as you're saying with the example of, say, you're making the spaghetti and everything with the spaghetti is happening on the cloud if it's if it's in the digital world. Why do so many other companies end up turning to Amazon and Microsoft and Google for the cloud services, for you know, their computational services? Why don't they do that themselves? As they develop this dependence on Amazon and Microsoft and Google, why don't they pull back?

SPEAKER_00 19:03

Well, actually, the book answers that question at different levels. Because one of the things that I've been identifying with these cloud giants, but also with other leading corporations, is that they build a sort of subordinated periphery of actors. They sit at the core of global capitalism and they have these satellite organizations. Some of them are, if you want, a semi-periphery or semi-core. And when we are talking about cloud giants, the semi-periphery or semi-core, of course, will include AI companies that struggle still to have their own business model, but exist because of the money that companies like Amazon, Microsoft, Google, and also Nvidia put into them on top of the traditional venture capitalists, and that also exists ultimately because they are providing knowledge and enabling the perpetuation of those that are sitting at the core. So this periphery, and I can expand on them later on, when it comes to the cloud, is populated by a lot of what the companies describe as their partners. And partners basically create specialize in one type of service. So you have companies that do a special type of software as a service, or you have other companies that instead of developing their own software as a service or an application of an AI model that will be sold as a service, and so on, you have other companies that operate as their sales force. Just like uh in other industries where uh the sales and customer service is outsourced. In part, Amazon, Microsoft, and Google do it themselves for big clients, but for the large majority, they use also these smaller companies that make a business out of uh selling more uh cloud migration, pushing more and more organizations to move to the cloud. This includes also a special role for big consultancy companies like the McKinsey, the Accenture, and so on. They also make a living out of uh convincing not only big companies in their case often, but also governments to migrate to the cloud and to start operating on the cloud and consume their own services like Accenture, IBM, they all offer services on the cloud of Amazon Migrants of Google. So all these partners that include many, many more, and there are thousands, in a way, they have no better choice than to join the cloud. Because by doing so, they make a bit of money. Of course, they do it in a way that is uh zero autonomous. They need to comply with the ways and the protocols decided by companies like Amazon to develop their own solutions. They need to follow the rules of an Amazon, Google, and so on when they are trying to sell these services and convince others to migrate to the cloud. And they will accept that they will make relatively less money, even if they are taking more of the risk when they are developing the new software. That then will be offered as a service on the cloud. And when the software is offered as a service on the cloud and it's sold, a piece of that will go to the cloud giants. So it replicates, in a way, the structure of the Android iOS. But the difference is that when we download the apps to our phones, the different apps are not necessarily interconnected. They don't need to be all connected to each other to work. Sometimes they are connected and it's a way to capture more of our data. But it but we can just use one app and that's fine. On the cloud, the services are all, again, need to be interconnected because they are all these Lego pieces that I was mentioning before, that once they are put one together with the other, plus your own code, you create your own solution, either one that you will be selling as a service or one that you will be directly using internally for your organization. For a company that is a partner, in the end of the day, this is a way to make money by taking most of the risks. But then we have the other part of this story, which is what happens with the big companies, what happens with governments. Why are they not just, and just one more thing about the smaller companies, the way in which they get hooked in the cloud is usually through the allegedly like savings that they will be making. Because if you're a startup, if you're and part of this story also uh explains why we have so many startups popping up all the time, you don't have the funding, you don't have the money to have your own data center, develop or purchase all the like a sort of bespoke set of services on top of which or through which you're going to create your own uh software model, whatever. The cloud offers you all that. So it's a supermarket where you can just go and you use, you pay as you go. And this is the other marketing strategy that has proven to be very successful. So, again, if you're a small company, you're not going to pay upfront a lot of money that you don't know if you're going to get back through your sales because you don't even have a product yet. So they offer you, even through cloud credits, the possibility to create your whole architecture inside the cloud. So cloud credits means basically you don't pay, you have like an opportunity to use services for free up to one point. But it's not that you were producing something and then you can just move with what you were doing and go to the next place. It's not like a warehouse. And then you move the different supplies that you still have there to the next warehouse and you continue producing once you have to pay or when the rent increases. Because the only way in which you can operate is through accessing all these black box services. So if you want to leave, you lose everything. So you get trapped, and the ways in which the structure of organizations works makes it self-reinforcing. Because the big get bigger and the smallers stay small and specialize. Even Amazon recommends the partners to specialize, to focus on a single product, to focus on a single industry to make money, which basically is a way of saying, don't you ever dare competing with me. You will make money, but always under my umbrella, always following my rules. So this is one side of the story of why they rule on top of other formally independent organizations. And this is part of what in the book I describe as these ways of controlling beyond ownership. Because even if the companies are not investing, are not venture capitalists investing in the smaller ones through the cloud, they will control anyway how every single organization is developing their own uh software and hardware and data solutions. But the other side of the story are, as I was uh introducing before, all these other organizations that are not native, if you want, from the tech sector. And they all migrate to the cloud. For them, also, these uh same advantages, if you want, are tempting. Uh, if you don't know if your business is going to always do well. And even if you trust your own business, you're a big company, you know you will be successful, but you also know that the world is extremely uncertain. And if there is anything, you have certainty that something will go wrong. And not necessarily again for your specific business, but it is like in the midst of an ecological breakdown with pandemics, with um geopolitical turmoil, all this may result in uh unexpected outcomes that may just make your business plummet and or have or close part of your business from one day to the next. So turning the fixed capital investment into a flexible one is very tempting, and you are willing to pay. You're willing to pay extra for that. So even if in the long run it's more expensive to be on the cloud than operating with your own data centers and developing all the architecture yourself or purchasing from different providers, independent providers, and keeping track and access to the code, all that in the long run is probably cheaper for big organizations. But still, this risk is one, one of the reasons why they prefer to migrate to the cloud. The other reasons are connected to what they get by being on the cloud. What are those technologies? And for big corporations, for companies like Nestle, Coca-Cola, IKEA, PepsiCo, but big pharma companies, automobile industry companies, these companies, more and more their leadership, their advantage is explained by developing intangibles, knowledge, narratives using digital technologies. And I will give you concrete examples. You're Coca-Cola and you need to continue building your brand. You cannot afford just doing brand building without data-driven technologies, without analyzing the data of your consumers, without using AI to try to predict, forecast the behavior of those that are parties in your products. If you're a pharma company, you're more and more compelled to adopt AI at every single step of the drug development process, even to push those that actually develop the drugs, which are the universities, the public research organizations, the biotechnology startups, to adopt AI as a method of invention. Because that seems to be the way to accelerate the discovery of new drugs. If you're an automobile industry leader, in the book I tell a bit of the story of Renault and Google and how Renault actually realized that it needed to have like frontier software inside the cars, that the cars more and more are becoming a sort of empty shell, and they need to fill inside that shill with digital technologies. So they wanted Google to develop a special application of Android for their cars. But Google was, uh because Google was ultimately in control, Google told Renault, okay, I will do it, but you need to migrate to my cloud. And Renault didn't want to do it because I guess they knew that they would end up being completely trapped, and not only in terms of prices, but also losing control over what technology they are consuming, but ended up agreeing because that was essential for them to get the digital technologies inside the cars. So and this is one of, I mean, on top of the flexibility connected to and the and the sort of paying for the insurance, I would say that on top of that, another reason is definitely the perpetuation ultimately of the leadership. What I would describe the perpetuation of these other leading companies as intellectual monopolies that are systematically turning knowledge, narratives, data into their assets. This more and more depends ultimately on digital technologies and through that on the cloud giants. But also, I was giving you before the example of the ERP, the enterprise resource planning, so the software that companies use inside the workplace to ultimately control workers and control the workers of their subcontractors. So if you are a leading corporation, if you are IKEA and you have thousands of global value chains to control, and you control them again beyond ownership, the way to do it, the way to centralize all the information, to know which supplier can do what, how they can you squeeze the most from each and every subcontractor, how to make sure that the workers of your subcontractor will deliver. And also with IKEA, IKEA has a franchising business. How to make sure that all the workers of the franchisees of all the stores are behaving in the same way, with the same tone of voice, uh, knowing the exact same information and talking to the customers in the same way. For all this, they use digital technologies. And all these digital technologies are a sort of mix of code developed internally by the leading corporations and a lot of black box services that they get on the cloud. And with AI, this process is fast-forwarding because AI is also being sold as a technique to control better and forecast better. So for corporate planning, it's very tempting to, if they were not already on the cloud, to migrate to the cloud to exercise acute control of all those that participate in what in the book I describe as spheres of control that basically include global value chains, price-chising models, platforms, you name it. So this is why they migrate. They migrate to the cloud, they become dependent ultimately because they get something in exchange. And the problem is that if we go further down in this hierarchy of power, although the ultimate rulers are Amazon, Microsoft, and Google in their power block, in their coalition with the US government, and then the next layer will be all these organizations, these leading corporations become dependent because that is a strategic way to get access to use actually without accessing the code, to be more precise, to the technologies that enable them to keep on leading and profiting and extracting value from those that are further below. So ultimately, it's at the expense of us as workers, us as the majority of the society. Although these are power relations among companies.

SPEAKER_01 32:17

I think that's a really important point to understand because it's easy to look at it and say, oh, they've been told that they will save money by going onto the cloud, but by recognizing that access to these services, access to this infrastructure, this network is part of what maintains their power and their dominance, you know, of the kind of non-tech companies in the 21st century economy, it gives us a new insight into how all this works, right? Into what can actually be done to try to rein these companies in as well, or at least to understand how they work. You know, you were talking about how small companies get hooked, you were talking about how big companies get hooked. But once all these companies are kind of hooked on the cloud, have developed this dependence to Amazon, Microsoft, and Google services, how does that benefit those companies, the the cloud giants themselves? How do they benefit from the fact that they have all of these companies now dependent on their service, on their cloud, on their platform?

SPEAKER_00 33:15

Well, the dependence, the how they benefit from this dependence, you need to think, of course, this dependence is an economic dependence on one hand side, so more money. And when people ask about how companies, again, like OpenAI or Cloud or Anthropic and so on exist, well, this is because there are other companies that are making tons of money. And although in the case of Amazon, its largest business in revenue is Amazon, it's a commerce marketplace, in terms of profits, Amazon Web Services is the largest one. So these companies profit a lot from these economic dependencies. But economic dependencies are not only the main benefit that they get. It's not that they are just capturing more and more money, which indeed they are. But precisely there is where the book, I think, contributes to go one step further, because we already know that these companies are the largest in the world. So definitely they profit a lot. The first thing that the book adds there to this picture is that they are profiting from co-created uh value. They are profiting from capturing this value, they are profiting from capturing not only data but also knowledge from all these other organizations that are developing services on their clouds. So basically, they are robbing us all. But on top of the economic dependence, this economic dependence mirrors to some extent forms of political dependence around the world. And they have become the most powerful actors. And the book is called The Rulers, because the point that I want to make and what I want to insist on is that it's not just that these companies are the new big companies in town. Ah, okay, before we had the rubber barons, and then in the 60s we could already see other big giant tech companies. IBM was big, and now it's no longer the queen of technology that it used to be. So eventually the quote-unquote market or antitrust authorities will take care of it, and it's everything is fine. The ways in which these companies exercise power, controlling beyond ownership and creating this sphere of dominated actors, results in parallel economies that cross borders. And these economies are ruled, are run, are dictated by authoritarian planners. Amazon, Microsoft, and Google are ruling our ways of operating, of behaving in spaces that go way beyond social media. I would argue that actually the fact that the founders are like billionaires, now trillionaires, or uh and how they control the public sphere that social media is today, are just the tip of an iceberg. And to really understand the roots of corporate power, we need to understand, we need to, in a way, trace, map all these power relations. To understand that although, of course, there needs to be a lot of research and a lot of investigation and a lot of scrutiny of companies like Palantir or companies like Elon Musk companies, we need to understand the role that each of these companies play in this big scheme, in this big structure of power, of hierarchies of power among different organizations. So I think because I think of the economy as necessarily entangled with the political and power relations, one could actually say they're one and the same. One could actually say, okay, but this in the end is just capitalism. But precisely because it's just capitalism to some extent, but also to others, it's still capitalism. But with if you want new features, we cannot just accept, okay, this is capitalism, this is big power. We need to understand concretely the ways in which this power is exercised. And I was saying one is economic power, this political power, but also there is an epistemic power. There is a control of even how the world thinks about these technologies. So it's not just that they control the actual production of digital technologies, the actual creation of new research, development, frontier AI models, and so on through their clouds and beyond, but also that they control how we think about AI, what AI is, what power means, what should be regulated, what shouldn't be regulated. And in the book, I speak about the idea of epistemic totalitarianism. Because they are through crafting narratives that go from regulation shouldn't stifle innovation to PDUM, like thinking that we are going to get extinct because of AI, all these crafted narratives ultimately shift the conversation away from corporate power, away from those that are actually deciding that this is the technology that we are getting. So, and this epistemic totalitarianism means that their AI is the AI that we got. And because AI is being used as a method of invention, I was mentioning before the drug development, but even in economics, we are pushed to adopt AI, use AI for something, to do research using AI as a method. So the more every discipline, the more every creative space, think of the creative industries, the more and more we adopt AI as a method of create as a method of creation, as a method of invention, the more they control how we even think, how we ask questions, how we answer questions. What is a relevant question and what is not. So it's not only about the need for pluralism, plural methods, plural ways of thinking, but it's also about how, through it and in an invisible way, these networks even reach how everyone conceives the world. And these days, with generative AI, their AI, the AI that they've been promoting and fostering, because the same model can be used for countless uses, you can adopt it for whatever you want. So it's ultimately cheaper for them to push everyone to adopt it. And again, they actually want to have a competitive market with generative AI models, because that further reinforces their power, their control of the whole ecosystem. If a company like OpenAI has to take care of all the competition and then is concerned about Claude and the other models from Anthropic, and is concerned about Elon Musk models and so on, it will never dare of becoming a cloud giant itself. So, in this process of controlling the whole economic and political space, they also control how we speak and how we think about AI. And I will just go back to, for instance, the idea of PIDUM. PIDUM, this discussion about are we going to, is AI going to destroy humanity and so on and so forth. Regardless of the answer, what we neglect when we are thinking of AI in those terms is that AI, although it, of course, there are certain things that the developers, the scientists get surprised of when they see how the models operate, there are decisions being made. The first decision was to produce LLMs. Like someone decided that we had to produce LLMs, and then that the LLMs had to be uh connected to us through interfaces like ChatGPT. And in the case of OpenAI and ChatGPT, this was Microsoft. This was Bill Gates pushing, like not even Microsoft abstractly, but the founder of Microsoft pushing OpenAI to develop uh ChatGPT. So there are decisions being made. There is someone pushing the accelerator, deciding that we are getting this technology. So if it's real, that it's so dangerous, so risky, and so on, why? Why do they keep developing it? And this also goes to all of those that are developing it. For the scientists and research, like and the people, engineers and so on, these narratives, the narratives are not only built for lay people, the narratives are also built for those developing the technology so that they don't feel accountable for their own decisions. Because the AI is doing it. The AI may have these consequences. So by controlling these and other narratives, I was mentioning the regulation shouldn't stifle innovation. And this one is very important because basically it's telling us a few things, starting from they are the innovators, and they should, and if you regulate them, you will not get the innovations. And why do innovation matters? Well, because the narrative of narratives, the ultimate narrative in this story, is that technology will bring progress. Technology will make us live all happily ever after, and so on. And again and again, it only takes one second to think of the history of technology, to realize that this is a lie. This doesn't hold water. This is not always the case. And it's not always the case for multiple reasons, not only because of uh paradigmatic examples of things that go wrong, like the atomic bomb, but also because it equally matters the impact that this technology is having in our workplaces, how AI is being used for control. Can we really think of AI as a technology that will bring progress? What do we mean by progress when this technology is used as an automatic weapon to kill people? Is killing people synonymous with progress? Is killing kids' capacity to think because they are using AI when they are at school instead of learning by themselves to think and criticize and even ask themselves the questions? Is really that progress? Is that going to bring progress? Or is it just that we are blinded by a promise and we are blinded by the idea, the abstract idea that if the economies grow, then that will be better for everyone. And we forget, again, that there are corporate decisions and political decisions, also, complicit political decisions being made that result in a technology that is controlling more and more spaces of our society.

SPEAKER_01 43:26

Yeah, I think that's really well said. And and really, you know, speaks to us about the much bigger picture of the control and the power that they're exerting beyond just, you know, a cloud and a product that they're offering. And I wanted to pick up on what you were saying there because you're talking a lot about the narrative. But there's also when you were talking about the intangibles, there's also this piece around the knowledge, right? Obviously, you are in academia, you have you have seen and kind of studied this, but how does their power then extend into the academic space and the research space to shape the actual kind of world? That is being done, the innovation that is being done, the kind of AI and technology research that is being done to ultimately serve their ends and kind of direct that work into the direction that they want it to go in.

SPEAKER_00 44:13

So let me go back to something that we were saying, and I will frame it within the other examples, which is control beyond ownership. This capacity to control how others are working, in the case of academia, how research is being done, even if you're not directly hiring those scholars, even if you're not directly telling them what to do. In today's capitalism, leading corporations do that all the time. And they do that with startups, as I was explaining before. They, in the case of cloud giants, they even control beyond ownership governments when they migrate to their clouds. They have the capacity to control beyond ownership other leading corporations. But if there is one place where this dynamic is particularly concerning, because it's also where we should, in principle, be building public alternatives. This is academia. And I would also add there in this story the open source community. So these are the, or used to be, the spaces where we were producing public knowledge. Public knowledge, and let me say something that perhaps is obvious, but more and more we need to emphasize that knowledge can be used by many at the same time without reducing our satisfaction. Actually, the other way around. We produce knowledge on the basis of existing knowledge. So we use knowledge to create more knowledge. So the more we broaden access to knowledge, which requires also broadening the tools to understand how knowledge is produced, the more knowledge we can collectively produce. And if we think that it's through knowledge that we can solve problems, and uh in today's world we have a lot of problems to solve, then we should actually be thinking critically about how to expand these public spaces of knowledge production. However, what has been happening and what in part explains this invisible capture by big tech companies, but also in the pharma sector by big pharma and so on and so forth, is that academia has been pushed to show economic value, has been also controlled, conquered by the narrative that I was describing before. The idea that the only knowledge that is valuable is the one that will create economic profits for some, that will bring, again, this is what will is supposed to be bringing uh progress for everyone. Instead of recognizing that by sharing knowledge, and think, for instance, of the case of the vaccine, if there is one case where this is obvious, is the COVID vaccine. We really needed to have that knowledge distributed very fast and without charging any price at all, because we needed to save as many lives as possible, and because all our possibility to leave the vaccine was all interconnected. And instead of distributing massively, opening up the knowledge, enabling public laboratories around the world to use the uh the different uh recipes, if you want, for the vaccine to produce vaccines massively at the scale that we needed, and to offer the vaccine to everyone, only the rich got the vaccine fast. And this was because they could pay for it. And this was because the vaccines, even if they were funded by the public sector, even if they were developed by universities like the Oxford vaccine, it ended up being in the hands of AstraZeneca, and AstraZeneca charging lower prices in Europe and higher prices in the rest of the world. So this same dynamic has been happening in academia in relation to tech. And when you think of an academic space that doesn't have sufficient funding, where if you want to do a project, you need to apply and wait if you get the grant. And if you get the grant, it may not be the money that you need. Think again, this for developing digital technologies, in particular frontier ones, AI models, you need a lot of people, a lot of resources in terms of infrastructure that the for processing the models, you also need data. So it's really, it becomes virtually impossible because the system was already prepared to push academics to feed into the predatory ecosystems of big tech companies. And this is what they have taken advantage from. They have been granting money, they have been offering cloud credits, they have also been massively co-authoring academic research with them and even hiring part-time scholars from top universities around the world to work for the companies. But of course, this hasn't resulted, if you want, into a virtuous ecosystem where all the knowledge is shared and so on. Of course not. Governments speak a lot about technology transfer, where there has been science and technology transfer, but from the public sector to the companies, not the other way around. The leading corporations, the big tech companies, in particular companies like Microsoft and Google have been extremely active in not only partnering here and there with MIT, Stanford, and the University of California, they have thousands of collaborations with scholars from around the world, Microsoft, even with a lot of scholars from China, where they co-develop research and then they capture and monetize that research. And of course, they are the only ones that keep the whole picture. And the same happens with open source. They put pieces like breadcrumbs of their proprietary software in open source, and they create a sort of periphery of dependent developers that start creating apps and other solutions on top of the open source piece of software. But then it is the cloud giant that will be developing the complementary services, the adjacent services, and will be capturing them all, usually moving even the open source service to the cloud as an open source service offered on the cloud, like Kubernetes and many more. So ultimately, all these are ways in which the companies manage knowledge. This knowledge management and in the long run, this way of planning the development of science and technology beyond ownership enables them to perpetuate themselves as intellectual monopolies and to then extract value from all those that migrate to their clouds. Because they will always offer on their clouds the latest solution, but not because they are the innovators. I insist regulation shouldn't stifle innovation. Okay, let's take it as it comes. You are not the innovator. If you're Amazon, Microsoft, and Google, you are only the co-creator of innovations that are developed with many more, but at the same time, you're appropriating that knowledge and you're steering the type of question that all those others are developing and producing. And even those that are not directly connected to the companies, because they end up setting the global research agendas. Those from peripheral regions, those that do not participate in the most known top AI conferences and so on, will be looking at what is going on there to decide their own research questions. So ultimately and indirectly, we'll be operating and creating knowledge that will further favor the research agenda of Amazon, Microsoft, Google, and in this case also in the AI agenda of Meta. Meta is also very active in trying to set and control these research agendas. So if you see the whole picture, it's a story where there seem to be many independent actors, thousands of actors creating AI, thousands of actors in digital ecosystems, multiple industries that seem to be operating independently. And we can just end up saying, okay, these are the inequalities of capitalism. Those that innovate profit more, and we need to live with it. This is not what is actually happening. What is actually happening is that there are certain companies that are planning beyond ownership the development of certain not only narratives, but also science, technology, in order to perpetrate themselves at the core in the position of power, which again is not only economic power, but it's also political power. And what we see when they lobby, what we see when they are shaping policies, is again the tip of the iceberg. Because the way they shape our lives and exercise political power is directly through their infrastructures, through their technologies. Technologies themselves are political. The fact that they are pushing for the development, as I was saying before, of certain uses of the technology, even in certain sectors, is not just about warfare, like tech for warfare is wrong, but if it's for healthcare is right. If the technologies that you are developing for healthcare are replacing the physicians, are used for just focusing on the symptoms and contribute to neglecting the root causes that explain why more people get sick, like people getting sick with hepatitis in Mexico, because they cannot drink water from the pipe, because all the water is being used by the data centers. So if we are not approaching healthcare with these broader mindsets, because the way in which we approach it is directly in this techno-solutionist way, then even in industries that could profit from other types of digital technologies, other specific types of models, what we end up getting is a smartwatch that tells us that we didn't watch uh walk sufficient steps today. What the book is about is to operate at this intermediate layer between the general things that we know are wrong, the ecological crisis is wrong. Uh the inequalities in our world, we know is wrong. If there's anything that should surprise us, it's not the level of inequality. Every time Oxfam comes with these like new data points, like the 12 richest uh billionaires own more than half uh percent of the population. Yeah, well, the problem that what really amazes me there is not the figure, but it's actually that nothing happens after the figure. So, and and these are things we know. And and the geopolitical uh turmoil, the way in which governments like the US government, the Chinese government control also beyond ownership, we know all this. And then we know that there are specific cases that are particularly wrong. But how to make sense of all this mess? So for me, this has been about understanding these networks of corporate power and explaining, describing them and trying to go to the root of what explains their corporate power. And this couldn't be explained without this systematic, massive capture of knowledge, public knowledge from academia and from open source.

SPEAKER_01 54:57

Yeah, I think very well said. And, you know, it's the capture of knowledge, but then also the shaping of knowledge and you know how we think and the types of things that we even think are possible, right? Is all kind of bound up in what you're saying. Um, you know, I have one more question for you before we uh end off our interview. And I feel like you have become, you know, a really important voice on digital sovereignty, certainly on the left, you know, in Europe and Latin America in particular. And, you know, you've laid out the scope of the problem that we face here in the dependence on these massive cloud giants, the services they offer, the way that they shape how we think about what is even possible. So, what then does a digital sovereignty agenda look like to push back against this, to try to do something else to actually challenge the power that these major companies have?

SPEAKER_00 55:48

Well, I actually, when when two years ago you and I and others were thinking about this and writing the reclaiming digital sovereignty report, who would have thought that digital sovereignty was going to become such a prominent topic in the political agenda? And I'm very happy that it's happening, but I'm also concerned because in this capture, in this process of setting the narratives, big tech companies are also trying to shape how we think about digital sovereignty and what digital sovereignty should be. And more and more we see ways of thinking of digital sovereignty associated with techno-nationalisms, with ways of thinking, okay, if the technology is produced inside my borders, then it's fine. It doesn't matter who produces it, for what purposes, it's inside my borders. Or a more sophisticated version of that is if it happens inside my borders and it's only done by citizens that are already like belonging to that border, to that country. So, and companies that belong to that country, then again it's fine. And what is missed, and what what I think in our report, and since then I've been I've been trying to advocate for, is to think of sovereignty as linked to governance and democracy. Sovereignty is about deciding who decides what technology we get and what technology we don't get, who decides that this is the AI that we need to adopt, who is deciding the data that is being harvested and what for what is being used, who is deciding where the data centers are going to be built and how many of those our world can cope with. So all these decisions today are being made behind closed doors in the most anti-democratic way possible. These are dictatorial decisions, this is authoritarian planning. These companies are deciding the future and the present of the world, and they are not even telling us that they are doing it. And they pretend to convince us that if they come and build a data center in our territory, that if they come and put open an RD laboratory in our territory, that will solve the problem. And I also use the example of the RD laboratory because I think that with the data centers, more and more we are raising awareness that these infrastructures remain completely under the control of the large tech companies, regardless of where they build them. And on top of that, because of the laws operating in the US and China, even their governments can get access to the data. But still, governments again and again follow for the trap of competitiveness and the promise that AI will bring competitiveness and growth. And because of that, they end up accepting that they need to take for granted that these companies are going to decide the present and future of technology. And I think that this is why digital sovereignty matters so much, and we need to fight for it, because it's a way to fight for democratizing the research agenda, for democratizing what technology we get, for what purposes, and who decides this. And here I think that although there are a lot of spaces where alternatives emerge, these alternatives remain like pockets, like disconnected proposals. And I will give you one example, a recent one, Tela Brazil. So Brazil developed this, one could say this uh alternative platform to Netflix, Amazon Prime, and the like that uh is broadcasting a lot of movies and local um entertainment films and series from Brazil, also historical ones that were, I don't know, like filmed decades ago, if not a century ago. So, and and this is amazing because it's a way to uh popularize Brazilian culture, keep it alive. And of course, Tele Brazil is uh has no subscription. Everyone in Brazil can access it and watch whatever they want. And this is great. But we cannot just think of an app, one single solution. We need to think of alternative ecosystems because otherwise, sooner or later, even the public solutions end up depending on infrastructures that are controlled by large tech companies. And here, of course, one can say, okay, but it's still nobody produces the GPUs. We we've spoken very little of NVIDIA. And in the book, what I try to describe also in the relation of between companies like NVIDIA and the cloud giants or TSMC and Cloud Giants is also to understand how, in that interplay, we have, if you want, forms of network collusion, where NVIDIA gets uh it's dependent on selling its GPUs to Amazon, Microsoft, and Google. Amazon, Microsoft, and Google try to be less dependent on NVIDIA GPUs by designing their own semiconductors. But all these companies ultimately, in this even as if even if asymmetric relations end up benefiting from our dependencies. But in the case of NVIDIA GPUs, it's totally feasible to purchase collectively NVIDIA GPUs and then build a cloud that is truly public, then build a digital space where the technologies can be shared, can be co-created, can be open source, can if there is a space for creating a market, that that market can operate under rules that are set from democratic institutions. And I think this is what is missing. It's missing an ecosystemic view. What is missing is to have political imagination and recognize that we cannot replace new forms of corporate power that control beyond ownership with single, independent, owned solutions. That will never be enough. We need to create ecosystems where we can really share knowledge for solving urgent problems. And this cannot just be done by the experts in, I don't know, open source communities or academic experts. This is also something important to bring to the table. This we are not just talking about developing new technologies, new science. We are talking about how we govern our societies. This is a political question. And because it's a political question, we need unions on the table. So workers need to be part of it from all the industries. We need civil society organizations, but we also need to decouple, keep away from the discussion forms of corporate power that directly or indirectly end up shaping the solution, even the alternatives that are trying to be built. So for me, digital sovereignty is about creating these alternative institutions where we can co-create knowledge that follows not the narrative of economic growth and promises, but actually from the outset that is developed, connected to the problems, connected to the different communities and thinking always within planetary boundaries.

SPEAKER_01 1:02:43

Yeah, I think very well said. And I feel like we're seeing the discussion move a bit, you know, like over time, whether it's the policy approach or, you know, the type of regulation, there is an evolution, but it feels like it needs to start going a lot faster in order to, you know, kind of face up to what we're seeing from Amazon, Microsoft, and Google and this tech industry and the power that they wield over our society to shape so much. Cecilia, it's been fantastic to get your insight on all of this to understand the scale of the problem that we face. Um, but if we don't understand it, we can't think of a way to push back against it. So I really appreciate you taking the time. Thanks so much.

SPEAKER_00 1:03:20

My pleasure. Always have pleasure.

SPEAKER_01 1:03:22

Cecilia Rickap is an associate professor at University College London and the author of The Rulers: Corporate Power in the Age of AI and the Cloud. Tech Won't Save Us is made in partnership with The Nation magazine and is hosted by me, Paris Marks. Production is by Kyla Houston. Tech Won's Save Us relies on the support of listeners like you to keep providing critical perspectives on the tech industry. You can join hundreds of other supporters by going to patreon.com slash tech won't save us and making a pledge of your own. Thanks for listening and make sure to come back next week.