Anthropic has been forced to abruptly disable its most powerful AI models for all users, after the US government ordered the company to suspend access to foreign nationals over national security concerns.
The export control directive targets Claude Fable 5 and Mythos 5 – versions of the company’s flagship Claude Mythos system – and follows the public release of Fable 5 just days ago. In a statement, Anthropic said the order meant it must “abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.”
“Anthropic abruptly disables its advanced AI models after US order restricting foreign access over national security fears.”
The company was not given specific details of the security concern, but said it understands the government believes there is a method of bypassing, or “jailbreaking”, a safeguard that would allow Fable 5 to be used in identifying software vulnerabilities. “We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities,” Anthropic said, adding that other publicly available models can discover the same flaws without needing a bypass.
Anthropic disagreed with the finding, saying: “We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.” The company had previously called for greater US oversight of AI, but said the government’s action on Friday did not follow “principles of fair and fact-based regulation.”
The order escalates a tense relationship between Anthropic and the Trump administration. The two are already involved in a separate lawsuit after Anthropic refused to allow the US military to use its AI tools for domestic surveillance and fully autonomous weapons systems. In response, the government placed Anthropic on a supply chain blacklist, set to take effect later this year.
When Anthropic first released Fable 5 publicly, it had touted various “safeguards” to prevent hacking – but acknowledged the risk. “Fable’s capabilities exceed those of any model we’ve ever made generally available,” the company said. The tool had been given privately to a handful of organisations in April for previewing, because it was so intelligent it could be dangerous. Some critics dismissed the “too powerful” warnings as marketing spin.
Finance, technology and government leaders had expressed concerns about the public rollout. Canadian finance minister François-Philippe Champagne told the BBC the attention was warranted because “it’s the unknown, unknown.” The European Union, which gained access to Mythos earlier in June after weeks of talks, said the development underscored “Europe’s need for technological sovereignty.” A European Commission spokesman, Thomas Regnier, said: “We take note of Anthropic’s statement and are assessing.”
Anthropic co-founder Jack Clark told BBC Newsnight last week that AI capabilities were expanding so rapidly the company thought there should be a way for the public to slow the technology’s advancement. “You want the option to be able to take your foot off the gas and put your foot on the brake,” he said.