What is the big shift you describe in The Techtonic Shift?

‘It’s both a shift in our worldview and in the way power is organised in the world. For a long time, technology was something that brought the world together and created cohesion within globalisation. It made distances smaller and made location seem less and less important.

Technology is now increasingly becoming a driving force behind the pulling apart of power blocs. Rather than only being a connecting force, technology is becoming ever more explicitly part of geopolitical competition. That is the techtonic shift I describe in the book.’

You call that old world the ‘digital Pangea’. What do you mean by that?

‘The idea of one big digital supercontinent in which location barely mattered anymore. Wherever you lived, you could listen to the same music, watch the same shows, and use products and services from all kinds of parts of the world. That gave rise to the idea that we were becoming part of the same digital world, to an ever greater degree.’

And that turns out to be an illusion, according to you?

‘For a large part, yes. That digital world is starting to break apart into different ideological and technological blocs. You increasingly see a clear American, Chinese and European bloc, each with its own ideas about technology, power and values. So that digital Pangea is starting to fall apart.’

“That digital Pangea is starting to fall apart.”

What role does AI play in that?

‘AI is currently the most relevant technology within that development. The special thing about AI is that you have to build certain values into the technology itself. You have to determine: what is good behaviour? What is correct? What is desirable? An AI system can’t escape those kinds of choices.’

Can you give an example of that?

‘Take the question of what an AI model should say about the status of Taiwan. The desired answer to that is very different in China than in the United States. Then you immediately see that there’s a particular worldview behind such a model.

Even with questions that seem purely factual, context plays a role. Which historical perspective do you use? Which events do you consider the starting point? AI ultimately normalises a particular worldview, at exactly the moment when those worldviews are being pulled further and further apart by geopolitical developments.’

“AI ultimately normalises a particular worldview, at exactly the moment when those worldviews are being pulled further and further apart by geopolitical developments.”

AI is often compared to a new arms race. Does that comparison hold up?

‘That comparison holds up partly. With nuclear weapons, you’re talking about technology developed primarily by states. There’s also no normal commercial use case behind nuclear weapons. With AI, that’s fundamentally different: a large part of the development takes place at private companies, with enormous commercial interests.’

And is that where the US and China differ fundamentally?

‘Yes, there’s also a philosophical difference behind that. In the United States you see a strong focus on AGI. Technology almost becomes a goal in itself there: we have to build ever more powerful AI and then see afterwards what we can do with it.’

How does China look at that?

‘China is much more pragmatic. There, the question is more: what is AI useful for? Can it be used for industrial policy, education, healthcare or surveillance? If an application contributes to the state’s objectives, it gets encouraged. Ultimately, the difference mostly comes down to who has control: in America the private sector, and in China, much more explicitly, the state.’

Can the American state still effectively control Big Tech?

‘That’s becoming increasingly difficult. Just because of the sheer size of these companies. Take NVIDIA: its market value is comparable to the GDP of a large European economy. That’s obviously a skewed economic comparison, but it does give an idea of the scale at which these companies operate.

But there’s something else going on too: states themselves have become increasingly dependent on these companies. Governments use their cloud infrastructure, software and other technological systems. That creates a strange situation in which a government wants to regulate a company, while at the same time being partly dependent on that very same company for its own functioning. That makes tough intervention more complicated.’

“A government wants to regulate a company, while at the same time being partly dependent on that very same company for its own functioning.”

Are tech companies slowly becoming a kind of state, then?

‘You certainly see that parts of classic state tasks are becoming dependent on Big Tech. Max Weber defined the state partly in terms of the monopoly on the legitimate use of force within a given territory. But even around security, defence and public infrastructure, tech companies are now playing an ever larger role.’

What is the ideology behind the big American tech companies?

‘Alongside commercial interests, there’s an ideological current around figures like Peter Thiel, who is often associated with accelerationism. An important starting point of that is that technology is objectively good, and that more technology is therefore ultimately better. From that line of thinking, technological development should be slowed down as little as possible, and AI should be developed as fast as possible.’

How does such an ideology change our view of human beings?

‘It becomes a very mathematical view of human beings, in which costs and benefits are constantly weighed against each other. That stands quite at odds with the post-war idea in the West that every individual has intrinsic value, regardless of how much economic or technological value someone produces.’

Can Europe still chart its own course within this power struggle?

‘Yes, it can, but it’s difficult. An important theme in my book is that societies need institutions through which they can keep a grip on technological change. Technology is constantly changing how society functions, while our institutions often only respond to that much later.’

“What are you going to do with technology to make people’s lives materially better? As far as I’m concerned, that’s what it should be about.”

What should Europe set against that?

‘Europe shouldn’t just think about how it can keep technology under control, but also about what it wants to achieve with it. We’re very good at the question: what do we want to ban or regulate? But at least as important is the question: what do we actually want to do with technology? There needs to be a much more positive political and societal narrative alongside that.’

Is our democracy able to withstand the pace at which technology is developing?

‘That worries me. One of the problems is that democracies don’t always function well under crisis conditions. In a crisis there’s less time for the normal processes of power and checks on power, while major decisions need to be made quickly.’

And technological change can reinforce such a crisis?

‘Power can then end up in places that aren’t institutionally designed for it at all. That’s why you need to think in advance about how you organise democratic oversight. Once such a crisis actually happens, you really don’t have time for that anymore. I think some assumptions underlying liberal democracy may simply have become outdated. A good example is the classic idea of the marketplace of ideas.’

But according to you, the classic marketplace of ideas no longer exists?

‘Not in that classic form. We’re increasingly living in an algorithmically-driven outrage factory. It’s not necessarily the best ideas that rise to the top, but the ideas that generate the most attention: the most emotional, polarising or provocative messages. Those are the messages people click on and react to.’

“We’re increasingly living in an algorithmically-driven outrage factory.”

What does that mean for democracy?

‘That part of our political system is based on a market mechanism for ideas that, in that form, no longer exists. That information market is largely driven by commercial actors and their algorithms. Within liberalism we’ve thought extensively about how to prevent a state from becoming too powerful, but much less about what happens when the exercise of power comes from the market.’

And what would you change around technology?

‘I would look much harder at addictive algorithms. They’re really poison for your democracy. They can be bad for mental health, reinforce polarisation and fuel the spread of disinformation. That’s something I’d intervene on much more aggressively.’

“Addictive algorithms are really poison for your democracy.”

What can Europe learn from China?

‘Mainly that you actively try to steer technological change. I’m not in favour of the Great Firewall or of the Chinese political model, but China did ask a fundamental question when the internet emerged: how do we keep a grip on this technological change?’

So what should Europe actually do?

‘Europe needs to make choices. Where do we want to be strong ourselves? Where are we dependent on others? And how do we want to position ourselves politically and economically relative to both the United States and China? Europe needs to think much more strategically about its position between those two power blocs.’

What should ultimately be the goal of AI?

‘That’s really the most important question for me: what are we going to do with it? Technology shouldn’t become a goal in itself. We need to think much more concretely about the problems we actually want to use AI for.

Are we going to use it to help solve the housing crisis? Can we tackle major causes of death? Can we make sure healthcare keeps up despite an ageing population? Those are ultimately much more interesting questions than simply whether we can build the most powerful model.

The most important question is ultimately very simple: what are you going to do with technology to make people’s lives materially better? As far as I’m concerned, that’s what it should be about.’

Interview by Bodiel Siebers

In this interview series, we speak with people from Amsterdam’s AI ecosystem who, through research, entrepreneurship, policy and practice, contribute to the development and responsible application of AI.