The Space of the World,Data Colonialism,and Social Contract: Reconstructing Digital Futures—A Dialogue with Nick Couldry

Theory and methodology: defining the crisis

However, the transformation driven by data is both deep and vast, and these conceptual frameworks are insufficient to capture it. It is almost as if our whole lives are being rendered into data. Working with Ulises Mejias, we call this datification, and it goes beyond what happens on social media. Every business is trying to capture data about the world and process it. So, we wondered what would be a better theoretical framework for giving us a sense of the size of this transformation? And we decided that the colonial framework was the best one.

Colonialism, as it developed in the West from the fifteenth century onward, was fundamentally about taking everything. When Spain and Portugal invaded the Americas, they discovered a new world to capture, one they had not known existed. We see a parallel with what happened from the 1990s onward: the internet suddenly offered a world that businesses could capture that did not exist before. Now, our emphasis is on the capture of data—that this is a continuation of the long history of colonialism in the West.

In some ways, Western colonialism ended in the 1960s, but in other ways, it continues and evolves. European extraction from the Americas and Africa fueled capitalism’s rise. Capitalism was only possible because of colonialism. What is happening with data is the latest stage of colonialism: data colonialism. The colonial logic of taking everything has not disappeared.

Generative AI provides a perfect example of this logic evolving. If we take the most influential account of recent capitalist change, surveillance capitalism, it focuses on assets extracted when we are under surveillance on platforms. But if you create a beautiful song and choose to put it online because you are proud of it, that is not “surveillance.” It is your decision to publish it. Yet it can still be taken as input to train AI models. AI takes everything. In that sense, colonialism, which took everything, every asset and resource, is a more accurate parallel for what is happening with data and AI than framing it only in terms of capitalism.

That period was marked by hostility to regulation and a belief that markets should be free to develop without oversight (very different from the situation in China). That climate encouraged the first mistake. We delegated to companies, whose primary goal is profit, not designing a better social world, the power to build not just one space or another, but what I call “the space of all possible spaces where we could be social.” Crucially, we allowed this to happen for a long time almost without any regulation. As a result, platforms became larger than the biggest countries. Facebook, for example, has roughly 3.2 billion users, more than India or China.

When any possible space of social interaction is routed through a platform, that higher-dimensional, overarching environment becomes the space of the world, the social world we end up living in when we spend so much of our time on social media, naturally wanting to connect with friends and family. And that is why I say the space of the world, not the space of social media.

Second, we allowed this very natural desire to connect to develop in a very unnatural setting: platforms designed to track us continuously and extract profit by incentivizing us to do more, to like, tag, follow, because attention can be sold to advertisers. This was a tremendous mistake that we made in allowing social media platforms to be designed this way.

Although the original mistake was made in the West, China has now perfected this model through TikTok and Douyin: tracking everything you do to predict what you will want to see, and profiting from it. This creates a vicious cycle. The platform’s design inherently encourages us to express preference, like this, not that, which naturally intensifies polarization. Business models then exploit this dynamic to keep us more engaged and more polarized, because polarization attracts advertisers.

Third, we allowed these platforms to scale as large as possible. A profit-maximizing platform wants the largest possible user base and the maximum possible interaction, because that is how it maximizes profit. But that is not how social life is built. If you host a party, in theory you could invite seven billion people, but it would be absurd. You invite twenty, maybe a hundred, because you have a room, you want people to talk, to look at each other, to enjoy a calm time together. Social life works better at smaller scales.

As a result, we built social media with a design fundamentally bad for social life, at a scale impossible for social life. That is why I needed the concept of “the space of the world” to name what went wrong at the level of space itself. The problems we discuss, mental health harms for young girls, political polarization, harassment and abuse, are symptoms of this deeper structural error. We designed the space wrongly, not only in particular platforms. That is the fundamental problem we need to correct.

Methodologically, what I tried to do was to combine my general theory of what went wrong with my reading of the past 20 years, a broadly social-theoretical reading. I am a social theorist, and I think the core problem lies in the design of space, not only in the design details of any particular platform.

At the same time, I tried to review as much relevant evidence as I could, from social psychology, sociology, some economics, and political communication, on specific issues such as polarization, declining trust in facts (and in the factuality of what we see on social media), and isolation: people becoming cut off from the normal social world. I reviewed that literature and then formed my own judgment about the balance of evidence. That is basically how I proceeded. Because it is a popular book, most of the detailed sourcing is in the footnotes, but if readers follow them, they can see the evidence behind the claims I make.

China’s place in the new framework

So yes, we need to listen to China, and it is important to have dialogue. It is not healthy to demonize China as a society that simply does not respect privacy. That is not true. There is evidence that Chinese people do care about privacy, and there are debates about privacy in China, though they take place within a different set of governmental values. I do not think we can have an intelligent debate about the future of social media, or AI, without engaging seriously with China’s policies.

What role China will play is difficult to predict, not least because US policy is currently hard to anticipate. But the future, in my view, depends on dialogue: listening carefully to what China gets right. It is important for people in the West to read Chinese policy documents closely and seriously, including AI action plans, Internet Plus, the social credit system and the reasons it was proposed, and the recent 5-year plan with its detailed proposal for rolling out AI across society. It is not what I would choose for Western societies, but it is an intelligent attempt to think through social order, its possibilities, weaknesses, and problems. My point is that sustained engagement with China is essential, and that is what I have been trying to do in my work for the last 15 years.

The United States clearly exercises colonial power through global platforms—Meta, Google, Microsoft—in very deep ways. China also has very large platforms with enormous influence. It is true that China has used the provision of technology, cloud services, and infrastructure to acquire influence across parts of Africa, Latin America, and Asia. In that sense, China should be aware of data colonialism. But this is not a story where only one side is “colonial” and the other is not. The point is that we need to be honest about the power dynamics and have more serious dialogue about what different ways forward could be.

At the moment, we are in a world where there is this massive conflict between the United States and China over world dominance. The United States is now in a defensive posture, deeply concerned by China’s stated ambition to become the world leader in AI by 2030. If we want to move beyond data colonialism, it would mean giving countries, and even local communities, more power to negotiate their own terms: control over digital infrastructure, the platforms they use, the contractual conditions, and the ways data can be collected and used. In other words, reversing data colonialism means decentralizing power, bringing it closer to communities, so that human connection is not treated simply as an asset to be exploited for profit by businesses or for power by states.

A positive direction would be using data to build common knowledge and strengthen the social fabric, something that may be happening in some parts of China and the West, but it is not the norm. The broader vision is to move away from treating data—and therefore our lives, insofar as they become data—as something to be extracted, toward treating data as a shared resource governed for common benefit.

Scholar responsibility and critical engagement

So, at a time of rapid change and real uncertainty about where we are going, and where we should go, the role of critical scholars is substantial. Our responsibility is to speak clearly, based on evidence and on decades of thinking about social change, about what is healthy for society. We need to explain the deepest problems, and the beginnings of solutions, in language that everyone can understand, without jargon. And that may mean disagreeing with governments. I disagree strongly with my own government’s policy on AI.

In fact, I think there is an even greater role for critical scholars today than 40 or 50 years ago, because technological change is now reshaping the fabric of society in profound ways. These changes are without precedent; perhaps nothing comparable has happened for 500 years. And communication scholars in particular have a special responsibility, because many of the deepest changes are happening through the redesign of everyday communication infrastructures.

As for the balance, this has been a debate in sociology for over 60 years. Since the 1960s, critical sociologists have argued that “pure objectivity” is a myth. Of course, we should try to minimize the influence of personal likes and dislikes, and we should aim to be as objective as we can. But our search for objectivity is still shaped by values, just as governments make value-based choices about what kind of society they want to build. Scholars also contribute to one kind of society rather than another. The key is to be honest and open about those values, not to hide them, while still being rigorous about evidence, clear about arguments, and as transparent as possible about what we find. I do not see a contradiction there.

Future research: Predatory AI and resistance

The core question is: what happens if we make the wrong deal, the wrong agreement, the wrong “social contract” around AI? By “social contract,” I do not mean it in the classic political-theory sense of an imagined agreement that legitimizes government. I mean the actual contract, the actual habits, norms, and practices that we are entering into over time in a regular way around AI. My fear is that we are drifting into the wrong social contract, and that is what I investigate in the book.

My worry is that if we get it wrong, more and more of our everyday thinking, planning, fact-finding, and even how we relate to the world, will be shaped by corporate sources, corporate habits, and corporate AI tools. AI as rapid calculation can be extremely useful, but the point is to harness that capacity for social goals and public good, not simply for profit. Otherwise, we repeat the mistake we made with social media: we allowed corporations to build our social world for their own benefit. With AI, the stakes are even higher, because we are talking about tools that shape the mind, how we think, and how our children will learn to think. Organizing that primarily around selling products and licenses is a dangerous model.

I am arguing that right now we are making the wrong social contract around AI because we are allowing corporate profit to be dominant. Instead, we need to put social values first.

A concrete example is education. We should be asking: what role should AI play in schools? Is it helpful for children to be encouraged to use ChatGPT all the time? Companies are already building these tools directly into everyday devices, encouraging students to scan an assignment and receive an instant starting point of an answer. If that becomes routine, why would a child do homework in the same way again?

In that sense, AI is redesigning what we might call “cognitive production”—how knowledge is made in society. If we do this the right way, it could be tremendous. We could enhance our own agency, using AI’s calculations to give us more information and possibilities. However, if we use it the wrong way, we will fundamentally undermine our own skills and those of future generations. We risk deskilling ourselves to the point where we become dependent on AI. While that scenario is very good for profits, it is very bad for the human mind.

Yet in practice, we are allowing these products to spread as fast as possible without asking the crucial questions first. Sweden, Take Sweden, for example, which has essentially declared access to ChatGPT a basic right for every citizen and child. This presupposes that AI is inherently beneficial, much like clean water. You would not let everyone have access to water unless you knew the water was safe. But do we know ChatGPT is good for people? We do not yet. We really do not.

We are taking massive risks without a serious debate about the potential costs to the individual and, consequently, to society. A robust society relies on the cognitive development of its members to form a collective intelligence. By rushing to adopt AI tools whose effects are unknown and unpredictable, we gamble with the very fabric of our social intelligence.

We need to decelerate and exercise much greater caution. In this regard, there is merit in one key aspect of the Chinese approach. Whatever other differences I may have with Chinese policy, the idea of framing what we are doing with technology in terms of social consequences—which has always been a Chinese approach—is correct and more useful than a simple pro-market approach. That is where we have to start, not from how to make maximum profit for AI developers. That is the wrong choice. I guarantee that.