Anthropic Abruptly Disables 2 Claude Models After US Bans Use by Foreign Nationals

Anthropic abruptly suspended all access to two of its artificial intelligence Claude models on June 12 after the United States ordered the tech startup to ban foreign nationals from using it.

The ban applied to all foreign nationals inside or outside of the United States, including Anthropic employees, according to statement released on June 12 by the San Francisco-based artificial intelligence company.

The directive by U.S. officials prompted Anthropic to go a step further, and to pause access to all customers using Claude Fable 5, a model that could be used for “ambitious coding projects,” and Mythos 5, a model designed for cybersecurity and biology research.

“Our understanding is that the government believes it has become aware of a method of bypassing, or ‘jailbreaking’ Fable 5,” the statement said.

The order from U.S. officials was issued early Friday evening, but did not provide specific details of concerns, per the company.

Anthropic believed the situation is a “misunderstanding” and is “working to restore access as soon as possible.”

“We suspect that perfect jailbreak resistance is not currently possible for any model provider,” the company stated. “Every safeguard used in the industry is vulnerable to non-universal jailbreaks (which can elicit some cyber information in specific circumstances), and it is likely that universal jailbreaks will eventually be found in the future. We stated this clearly when we released Fable 5.”

Before the company shut down access to the highly advanced models on Friday, Claude Fable 5 was available to the public and Claude Mythos 5 was only available to “a small group of vetted partners” on June 9.

“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 AI company added. “If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.”

The company reiterated that access to all other Anthropic models will not be impacted; the unaffected Claude models include Opus, Sonnet, and Haiku.

Friday’s directive was the latest chapter in the Trump administration’s approach to the rapidly growing artificial intelligence industry.

On June 5, Trump issued a memo aimed at speeding up the development and use of artificial intelligence for national security purposes, but he did not mention Anthropic’s participation, as the AI firm and the Pentagon have been at odds since the spring after the AI company refused to grant government officials unrestricted access to its Claude models. 

At the time, the company was concerned that its technology would be used for mass domestic surveillance or fully autonomous weapons, which the Pentagon denied.

The government responded by placing Anthropic on a supply chain blacklist.

Aldgra Fredly and Reuters contributed to this report.