Anthropic’s decision to shut off access to its most advanced artificial intelligence models raised alarm bells for European leaders already worried about over-dependence on foreign partners. 

In recent years, the EU has sounded alarms about too much dependence on Russian energy and Chinese manufacturing. Now, the decision by U.S. artificial intelligence leader Anthropic to suspend access to its Fable 5 and Mythos 5 systems has leaders worried about overreliance on U.S. technologies. 

Earlier this month, the White House ordered Anthropic — owner of artificial intelligence assistant Claude — to cut off access to its highest level systems for “any foreign national, whether inside or outside the United States, over security issues. Anthropic responded by cutting off access to the models for everyone. 

Fable 5 is the consumer version of Anthropic’s AI agent, while Mythos 5, known for its ability to detect cybersecurity vulnerabilities, is the version that was accessible only to a select network of digital infrastructure companies. 

Worries that Trump could "pull the plug"

The development comes after high-level worries that President Donald Trump could “pull the plug” on large swaths of the Internet in Europe.

“It is no longer reasonable to assume that we can totally rely on our American partner,” Matthias Ecke, a German member of the European Parliament said last year. It also comes amid debate within Europe over the landmark AI Act designed to regulate artificial intelligence. 

Even Pope Leo XIV has entered the debate, issuing his first encyclical, Magnifica Humanitas (Magnificent Humanity) in May, calling for artificial intelligence to be reined in to assure that it “serves humanity.” The sprawling document was unveiled last month by the pontiff alongside Christoper Olah, Anthropic’s co-founder.

"Preventing AI from dominating humanity," Vatican says

“Artificial intelligence now demands to be disarmed, freed from logic that turns it into an instrument of domination, exclusion, and death,” Leo said. “To disarm does not mean rejecting technology but preventing it from dominating humanity.”

But European tech startups and their advocates have called for the implementation of the AI Act to be delayed or for the act itself to be weakened, arguing that it makes it too hard for them to compete against rivals in the U.S. and China. 

“The EU’s AI Act risks creating a fragmented, unpredictable regulatory environment that will undermine innovation, discourage investment, and ultimately leave Europe behind in the global AI race,” read a 2025 letter signed by dozens of European AI founders and investors. Rules are necessary, the letter said, “but don’t regulate us into the ground.”

Now, those voices are expected to be taken more seriously. The AI Act now appears almost certain to be “watered down” and Europe’s lack of competitiveness in the sector is now set to become a central priority at the Group of Seven (G-7) leaders’ summit set to get underway in France on Monday. 

But experts say the AI sector in Europe will continue to suffer from insufficient investment, high energy cost, a complex regulatory environment, and the absence of national champions. 

A new fictionalized scenario called Europe 2031 earning headlines in European media points to a future where “over-regulation and a lack of ambition” mean that Europe is “left powerless as AI denominates defense, cybersecurity, and geopolitics."

Unlock unlimited access