How AI Is Reshaping Global Politics

A senior European official made a statement that stuck with me a few months ago in a quiet corner of a defense conference in Brussels. He claimed that no longer creating policy was the most difficult aspect of his work. It was telling ministers that a Taiwanese chip factory was more important than a Russian missile silo. When he said that, he didn’t really laugh.

In a nutshell, that is where global politics have strayed. Power was measured in tanks, trade routes, and treaties for the majority of the last century. Beneath all of this is now a more subdued ledger that is measured in compute, data centers, and the names of businesses that the majority of voters were unable to identify from a lineup. Even though the speeches in capital cities continue to sound the same, it’s difficult to watch this unfold without feeling as though something fundamental has changed.

The American strategy has been largely market-driven and hands-off, depending on a few well-funded labs and private companies like Google and OpenAI to keep the nation ahead. China has adopted a different approach, integrating AI directly into its industrial base and treating it more like infrastructure than a separate industry. This is a polite way of saying Beijing wants AI in every factory, port, and rail line it owns. The Tony Blair Institute recently referred to this as a shift from technological development to technological diffusion. It is another matter entirely if that tactic is successful. China has an abundance of scale, which is helpful, but scale is not the same as breakthrough.

Speaking with those in the field gives the impression that diplomats are discreetly catching up. These days, AI is used in intelligence operations, foreign policy modeling, and even negotiation preparation. Algorithms identify deepfake footage before it appears on the evening news, scan satellite photos for troop movements, and keep an eye on social media for early indications of unrest. A portion of this is actually helpful. To be honest, some of it appears to be governments spending money on instruments they don’t fully comprehend. It is possible for both to be true simultaneously.

The more difficult questions are located in the wiring, lower down. No treaty has yet to address the concerns raised by autonomous weapons, which are sometimes referred to as “killer robots” by those who obviously want them outlawed. I can think of at least three nations where deepfakes have already interfered with elections without doing any research. Furthermore, some academics are rightly referring to the growing digital divide between AI-rich and AI-poor countries as “digital colonialism.” Smaller states are being asked to choose sides in a contest in which the majority of them had no involvement at all.

The fact that the rules are still unclear is startling. China has its 2030 plan, the US has executive orders, the EU has its AI Act, and the UN continues to convene experts in Geneva rooms. These efforts don’t quite match. There’s a chance that a worldwide framework will eventually develop, just as it did with regard to chemical agents and nuclear weapons. This technology might also advance more quickly than the diplomats pursuing it, which would be a first in terms of degree rather than kind.

For the time being, those in charge of ministries are acquiring new vocabulary, sometimes with reluctance. Calculate. deduction. models of foundations. Last winter, an aide told me that he had just concluded a briefing outside a government building in London, during which half the audience asked what GPU stood for. The same group was discussing export restrictions on those same chips six months later. Nobody in Washington, Beijing, or anywhere in between seems completely certain of how this will end, and there is a steep learning curve and real stakes.