This weekend, the American flagship of artificial intelligence, Anthropic, was forced by its own government to cut access to its most powerful models, three days after their launch, Washington invoking a risk to national security in an unprecedented decision.
Anthropic announced it would suspend its two new models, its very restricted Mythos 5 and its bridled-for-the-public version, Fable 5, in order to comply with an injunction issued in the evening.
Washington ordered, under export controls, to cut access to these models for “any foreign national, inside or outside the United States,” including “foreign employees” of Anthropic, according to the company’s press release.
Unable to filter its users, Anthropic, already at odds with the Trump administration, announced it would have to “brutally deactivate” the two models for all of its customers.
According to the American media Axios, the directive originated from Commerce Secretary Howard Lutnick. He reportedly acted after learning that a company had managed to bypass the safeguards put in place for these models, reputed to detect and exploit cybersecurity flaws with unprecedented speed and acuity
When contacted by the AFP, the Department of Commerce did not respond.
“We contest that the discovery of a potential circumvention justifies pulling a commercial model deployed to hundreds of millions of people,” wrote Anthropic, which wants to view the affair “as a misunderstanding.”
“If this standard were applied to the entire sector, we believe it would essentially halt all new deployments of leading-edge AI models,” the company adds, at the forefront of the global competition against OpenAI, Google and the Chinese DeepSeek.
Anthropic states it is working to restore “as soon as possible” access to these cutting-edge models, the others remaining operational.
The company has long campaigned for public oversight of AI: in an essay published on Wednesday, its chief executive Dario Amodei argued for a regime of mandatory audits of the most powerful models, inspired by civil aviation, which would give the government the power to block deployment.
But this must be done “within the framework of a transparent legal procedure, fair, clear and based on technical facts,” recalls Anthropic. Friday’s directive “does not respect these principles,” the company denounces.
