US blocks foreign access to Anthropic’s top AI models over security
5 verified4 unconfirmed1 contested
Anthropic has abruptly disabled access to its two most advanced AI models, Claude Fable 5 and Claude Mythos 5, after the U.S. government issued an export control directive citing national security concerns. The order, which applies to all foreign nationals inside or outside the United States—including non-U.S. Anthropic employees—prompted the company to block access for all customers to ensure compliance. The government cited a newly discovered method of jailbreaking the models, though Anthropic stated its own tests found only minor, already-known vulnerabilities. The ban marks a major escalation in tensions between the company and the Trump administration, driven partly by Anthropic’s earlier refusal to allow military use of its AI for surveillance and autonomous weapons. The directive is seen as a potential setback for Anthropic’s plans for an initial public offering later this year at a valuation approaching $1 trillion. U.S. export controls have historically targeted AI hardware, making this a significant shift toward limiting foreign access to AI models themselves.
What’s verified
The U.S. government issued an export control directive citing national security to suspend foreign national access to Claude Fable 5 and Claude Mythos 5.
Anthropic disabled the models for all customers to comply with the order.
The government believes it has found a method to jailbreak Fable 5, according to the directive.
The order applies to foreign nationals both inside and outside the United States, including foreign-born Anthropic employees.
The ban threatens Anthropic’s planned IPO, which could value the company at up to $1 trillion.
Where accounts differ
The sources differ on the severity of the jailbreak vulnerability. One source states that reports of a jailbreak method are “not yet confirmed but likely true.” Another source reports that Anthropic said its own tests found only “a small number of already-known bugs” that were “minor vulnerabilities,” and that the company disagreed with the government’s conclusion that the findings should cause a recall. No other factual contradictions were identified across sources.
Not yet confirmed
Only one source reports that the Pentagon added Anthropic to a supply chain blacklist set to take effect later this year.
Only one source reports that President Donald Trump signed an executive order earlier this month requiring federal vetting of advanced AI systems for up to a month before public release.
Only one source details that Anthropic refused to allow the U.S. military to use its AI models for domestic surveillance and fully autonomous weapons systems.
Neither source specifies exactly how the jailbreak method works or whether it has been independently verified.
Key figures
Anthropic (AI company)
Claude Fable 5 (AI model)
Claude Mythos 5 (AI model)
Donald Trump (U.S. President)
Sources: marginalrevolution.com, dw.com
