Anthropic pulls Claude Fable 5 after US government order citing national security

Anthropic announced it is disabling two newly released AI models, Claude Fable 5 and Mythos 5, after receiving an export control directive from the US government on Friday afternoon. The order, citing national security concerns, demanded that Anthropic suspend access 'to any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.' To ensure compliance, the company removed access for all customers. The government indicated it had discovered a method of jailbreaking Fable 5 to identify software vulnerabilities, but Anthropic claims the technique is narrow, only reveals previously known minor bugs, and can be replicated by other publicly available models without a bypass. Anthropic had released Fable 5 earlier this week with safeguards to prevent responses about cybersecurity, biology, and chemistry, and had done so in collaboration with the US government.

This incident is the latest flashpoint between Anthropic and the Trump administration. Earlier this year, the Department of Defense labeled Anthropic a 'supply chain risk' after the company tried to impose red lines on military use of its technology, effectively barring government agencies and contractors from using it. Anthropic responded by filing lawsuits against the administration. CEO Dario Amodei has expressed support for a fair and transparent government process to block unsafe AI models, but argues that this action lacks transparency and does not adhere to those principles. The swift shutdown of the models raises questions about how the government will handle future AI releases and whether other companies might face similar orders.