EU AI law gets new support after warnings from AI researchers

The doomsday scenarios have been raining heavily lately. Fugitive AI agents who single-handedly stage massive hacker attacks. Malicious code that can pose an existential threat to all of humanity. And then the pleas of the AI ​​engineers to slow down the rapid development.

The warning that caused the most shivers recently came from Jacob Coxon, a renegade researcher at AI giant Anthropic.

"The people who build AI really believe that the technology could kill us all before the end of the decade," he wrote on X.

In the US, where most of the tech companies are based, the country's president has dismissed the warnings as nonsense.

"The only protection you need against AI is a strong and smart president with a high IQ," declared Donald Trump recently.

In Europe, the tone is different. The world's first comprehensive AI legislation, the AI ​​Act, was already launched here two years ago. The purpose of the controversial regulation was to protect Europeans from AI variants and ensure that the services provided in the Union are safe and ethical.

European Commission President Ursula von der Leyen described the law as "historic" and said it took "European values ​​into a new era".

However, the criticism was not long in coming. Typical of the EU, complained representatives from the tech giants, who argued that the rules would throw Europe back into the analogue age.

The image of a plastic bottle with the cork stuck in a flap was meant to symbolize the EU's eagerness to regulate already existing technology instead of promoting new innovations.

"If Europe continues on its current course, a chance that only comes once in a generation is missed," wrote the Swedish Spotify founder Daniel Ek in an appeal, together with Facebook's Mark Zuckerberg, in The Economist newspaper two years ago.

It was not made better by the fact that the former ECB chief Mario Draghi warned at the same time that Europe had already lost the race for the new technology and had been overtaken by China and the USA.

The EU responded to the criticism as it does so often these days: with so-called omnibus packages, which is EU lingo for reducing the regulatory burden on companies. Among other things, the companies are given longer time to fulfill certain requirements in the legislation. New simplifications regarding GDPR and data management are on g.

But after the last few weeks' alarm from within the tech industry, the EU's rules have come into a completely new light. As those who build the models plead for regulation, the Brussels bureaucracy suddenly appears prescient.

In her speech to the Union last week, Ursula von der Leyen flagged that she will invite representatives of the leading AI laboratories to a discussion on slowing development.

Maybe the AI ​​team didn't get it exactly right from the start – it rarely does. But while the US relies on Donald Trump's IQ, Europe has at least tried to think one step further.

Read more: Karl Dalén: This is why Trump refuses to listen to the warnings about AI