Commerce Dept. Shuts Down Anthropic's Fable 5 Over Jailbreak

Unprecedented shutdown affects all users

The U.S. Commerce Department invoked national security export controls late Friday to force Anthropic to completely disable its newest AI models, Fable 5 and Mythos 5. The directive bars distribution to any foreign national, including those physically present in the United States and even Anthropic's own non-citizen employees.

Because the restriction extends to foreign nationals working within U.S. borders, Anthropic determined it had no technical means to comply while keeping the models operational for anyone. The company shut down access for all users. Its less powerful Claude models, including Claude Opus 4.8, remain available.

According to Anthropic, the Commerce Department letter arrived at 5:21 pm Eastern Time and provided no specific details about the national security concern. Company officials later learned the decision stemmed from government awareness of a technique that bypasses Fable 5's safeguards—protections designed to prevent users from accessing the full cybersecurity capabilities of Mythos, the underlying model.

Why it matters

This marks the first time the U.S. government has used export controls to force an immediate, complete shutdown of a commercial AI model already deployed to hundreds of millions of users. The action establishes a precedent that could reshape how frontier AI companies approach model releases and safety claims, while raising questions about whether national security tools are being wielded for political purposes against a company already in conflict with the administration.

Company disputes narrow jailbreak justification

Anthropic characterized the jailbreak as narrow—capable of unlocking Mythos's cybersecurity abilities in only one specific instance rather than defeating all safeguards universally. The company also noted that similar jailbreaks could likely work on other publicly available models, including OpenAI's GPT-5.5, which face no comparable restrictions.

"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," Anthropic wrote. "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 called for a statutory process that is "transparent, fair, clear, and grounded in technical facts," arguing this action fails those principles.

Political context complicates interpretation

The directive arrives amid escalating tensions between Anthropic and the Trump administration. In February, President Trump ordered federal agencies to stop using Anthropic's models after the company refused Pentagon contract terms that would allow its AI to be used "for any lawful purpose." Anthropic had sought exemptions for autonomous weapons systems and mass domestic surveillance.

The Pentagon subsequently declared Anthropic a "supply chain risk" in March, prohibiting military and defense contractor use. Anthropic is challenging that designation in federal court. Key Trump advisors, including former AI czar David Sacks, have publicly criticized the company as "woke" and accused it of "regulatory capture."

Policy experts expressed confusion about whether the action represents targeted enforcement against Anthropic or broader national security hawkishness. Dean Ball, an AI policy expert, called it "simply cartoonish" and noted the contradiction with the administration's willingness to export advanced AI chips to China.

Others argued Anthropic invited scrutiny by emphasizing Mythos's dangers in its own marketing. "If you describe your product as a munition in every press release, eventually a government takes you at your word," cybersecurity researcher Peter Girnus observed.

These details were first reported by Fortune.