Many years ago, I was part of an Anglo-German debating club that met once a year. The routine was always the same. The economists in the group would escape to one room, the foreign policy experts to another. At the end of the conference, in a joint session, a “rapporteur” from each group would inform the other of their deliberations. Each side would listen politely, pretending to be interested in what the others had to say, their minds already wandering to the train journey home.
Now imagine, for a moment, if you added a third group to the mix — tech geeks and AI specialists. The macroeconomists would not even pretend to be polite. They think of AI as a bubble and feel mortally offended by the notion of cryptocurrency, which challenges everything they have ever been taught about money. Foreign policy people wouldn’t care much either, except about the narrow issue of AI-based weapons systems.
That would be a mistake. Technology in general, and AI in particular, will help determine the world’s geopolitical makeup for decades to come. More than that, and as my theoretical conference implies, it’s the intercession of AI with economics and foreign affairs that is truly interesting.
AI, after all, will likely lead to the biggest productivity shock of our lifetimes. US banks have already warned of massive job cuts in the years ahead. Industrial companies will be next. Through its effects on the economy, AI will change geopolitical relationships too. If you are truly interested in geopolitical power, then, you should spend some quality time understanding how this technology works, and what it can and cannot do.
This is what I did a little over a year ago, when I set aside a few months of my life to create an AI model that produced a European news service. Twice a day, it would scour through hundreds of websites and RSS feeds across dozens of European countries, translate the headlines, select the stories, summarise them, and produce a more-or-less readable rendition of events. A typical example here might be a political crisis in a minor EU state.
To be clear, my toy model is not ready for commercial rollout. It was a toy project, basically done for fun. But it did teach me a few important things. For starters, AI is not about intelligence at all — but about the ability to do routine jobs at scale. We think of chess grandmasters as super-intelligent, but really their edge over the rest of us is their ability to memorise vast amounts of chess moves, a skill computers already started to match in the Nineties.
With that in mind, AI will never replace investigative journalism, or original commentary. But it likely will replace journalistic drudgery, as my experiment demonstrated. Without AI, the news service would have taken two or three junior journalists to create. With it, however, and based on my experience in the legacy media, I would estimate that AI can replace some 60-80% of reporting jobs. Some newspapers are so predictable they could be run entirely by bots — and readers are unlikely to even notice the change.
The point here is that AI is not creative. It cannot produce true art or insight. It is exceedingly good at producing stuff that is pleasing to the eye, yet is unoriginal almost by definition. Imagine you asked 100 wedding guests to take photos of the big day on their phones. There may be some good ones, but it wouldn’t be the mobile phones actually taking the pictures. Rather, it would be people: holding cameras powered by AI. Much of the power of this technology comes through how it interacts with humans.
What has this got to do with geopolitics? That AI’s usefulness depends on how people — industrialists, politicians, nations — exploit it And here it is clear that the countries most keen to adopt AI in civilian life, like the US, China and India, will increase their global impact. The region that will most resist AI, and indeed already is, is Europe. One of the great European delusions has been that the EU can assert its power through regulation, especially the regulation of data and AI. In the meantime, their industries are losing their competitive edge precisely because they are not adopting AI when others are.
“One of the great European delusions has been that the EU can assert its power through regulation”
China is ahead of its rivals in adopting AI production technologies. Beijing and Washington are also advancing in developing self-driving cars, exactly the areas from which future growth will come. This is especially frustrating given the broader context. All European countries are struggling with deindustrialisation, the loss of tax revenues, and a lack of fiscal space. France is teetering on the edge of a financial crisis. Even the Germans had to borrow money to pay for their defence spending, and every European economy is too weak for the geopolitical investments they must make.
More than that, European countries have suffered falling productivity growth since around the time of the global financial crisis. A positive productivity shock, of the kind of mass adoption of AI will produce, would spur economic growth and private sector investment. It would lead to higher tax revenues, allowing governments to invest in public sector infrastructure and security.
AI may even arrest the rise of radical parties. What the far-Left and far-Right have in common is that they appeal to nostalgia, to a better time in history. Yet by rejecting new technologies, these politicians seek merely to slice up the existing economic pie. All the while, the liberal, multilateral order is fading precisely because it failed to achieve the one thing it was supposedly good at: growing the economy.
Whatever my foreign policy colleagues might imagine, AI matters to geopolitics not because of weaponry — but in our ability to afford it and other technologies that determine geopolitical strength.
The US still has the overall edge in AI, but the big benefit does not come from making it. It comes from using it. This is why, in theory, Europe still has a chance, despite having missed out on the first phase of AI development. But Europe will lose this battle too. The mindset that stopped the continent from making this stuff is the same that will stop them it using it. The Europeans obsess about data protection; the Americans and Chinese get excited about data startups.
Even so, you would be mistaken to think that Europe’s problem is primarily one of politicians trying to overregulate the economy.
The bigger problem is the companies themselves, which double down on doomed 20th-century strategies. Last week, Mercedes relaunched its prestigious S-Class model: the petrol-driven version. The very same day, Tesla announced it is getting out of the car business altogether, to focus totally on the development of AI. Elon Musk has concluded, correctly in my view, that the big automotive money in future will come not from making cars, but through AI and autonomous driving. The Europeans are absolutely nowhere in this race, even as the AI-based technology behind the self-driving car will be similar to that of the self-driving tank, or indeed the self-piloting drone.
To get good at this, you need a lot of energy, and China has the world’s largest renewable energy programme. Its electricity supply projections are off-the-charts, literally, if you compare them to the US and Europe. Cheap energy, willingness to adopt AI, and growing populations are the main determinants of geopolitical power. China’s demographics may be tricky — but at least it has the other two.
Among the Europeans, one of the countries that could do better than others is the UK. Much of the EU’s anti-tech legislation, from the AI Act to the Digital Services Act, came after Brexit. Unfortunately, the UK still follows the EU’s General Directive for Data Protection (GDPR). When the EU drafted the GDPR in the 2010s, it aimed to stop unsolicited emails or nuisance calls. Yet in its eagerness to get rid of them, Brussels also destroyed the viability of data businesses in general. To give just one example, GDPR is one of the reasons the Europeans are still not collecting data from self-driving cars.
Another country well-positioned to enjoy productivity gains through AI is Russia: because of its vast energy resources. That would surely qualify as an event of geopolitical significance for Western Europeans who fret about their security.
The losers of this new tech race will be those unwilling to change who are prone to compartmentalised thinking — similar to what happened in our conference of economists and foreign policy experts. To put it simply, you do’t want to leave geopolitics to the experts. If you are a leader intent on achieving strategic greatness, rather, my advice would be to surround yourself with tech geeks.
