Anthropic's top AI models, Mythos and Fable, have been taken offline after the Trump administration imposed stringent export controls, citing national security concerns. The freeze came after Amazon CEO Andy Jassy called Treasury Secretary Scott Bessent on Thursday, warning that the models could be jailbroken, according to Axios. An administration official claimed Anthropic “knew a jailbreak could happen and chose to distribute it anyway.”
At the heart of the dispute is a breakdown in communication between the company and the White House. Sources told Axios that Anthropic failed to “honor” a recent cyber executive order, and its leadership did not take the matter seriously enough. “Everybody said Anthropic was a bad actor. Some of us said it was time to give them a chance. Now those people are questioning that. They screwed us,” an administration official said.
The company, however, maintains it received explicit government approval to deploy Fable. The clash highlights the vulnerability of leading AI firms to sudden regulatory action. Crypto Briefing reports that the incident may accelerate the shift toward decentralized AI systems as a hedge against centralized oversight.
Export restrictions now prevent Anthropic from using or distributing the affected models globally. The company is actively seeking a deal to lift those controls, though no timeline for resolution has been announced. The freeze impacts both U.S. and international customers who relied on Mythos and Fable.
Some industry observers argue the administration's response was disproportionate and that clear approval processes could have avoided the standoff. The episode underscores the high stakes as governments grapple with how to govern the world's most consequential technology.