What happened
Anthropic disclosed in a blog post on Friday that it has taken its newly launched AI model, Claude Fable 5, out of users' hands. According to the company, the government's order directed it to suspend access for "any foreign national, whether inside or outside the United States, including foreign national Anthropic employees." To be sure it complied, the company went further and removed access for every one of its customers. The episode is without precedent and marks the newest flashpoint in a deepening standoff between Anthropic and the Trump administration.
What the letter said
The company says the government's letter arrived at 5:21 pm ET. "The letter did not provide specific details of its national security concern," Anthropic wrote.
The firm added: "Our understanding is that the government believes it has become aware of a method of bypassing, or 'jailbreaking' Fable 5. We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly-available models are able to discover them as well without requiring a bypass."
Anthropic's defense
In the post, the company maintained that it has built strong safeguards to lower the chance of Claude Fable 5 being misused. Anthropic also argued that the jailbreak the government identified was narrow in scope and would not have made an attacker meaningfully more dangerous than they would have been using some other AI model.
"To date, the government has only given us verbal evidence of a potential narrow, non-universal jailbreak, which essentially consists of asking the model to read a specific codebase and fix any software flaws," the company said. "Our understanding is that one potential jailbreak was shared with the government."
What Fable 5 actually is
Anthropic publicly released Claude Fable 5 on Tuesday. It is a version of the company's Mythos AI model fitted with safeguards that stop it from answering questions about cybersecurity, biology, and chemistry. The company says it carried out the public release in collaboration with the US government. Before that, in April, the Mythos Preview AI model had gone out in a limited rollout. The idea was to let companies and organizations tap its powerful cybersecurity capabilities to strengthen their own defenses, while easing fears that bad actors could exploit the technology to build potent hacking tools.
A relationship already under strain
Earlier this year, Trump's Department of Defense branded Anthropic a "supply chain risk" after the Claude-maker tried to set red lines on how the US military could use its technology. That label effectively shut government agencies and contractors out from using Anthropic's tools. The company hit back by filing lawsuits against the Trump administration.
The company's objection
Anthropic CEO Dario Amodei wrote in a policy essay earlier this week that he and the company back a fair, structured, and transparent government process capable of blocking the release of unsafe AI models. In Friday's blog post, Anthropic argued that "this action does not adhere to those principles."
Spokespeople for the White House and the US Commerce Department did not immediately respond to a request for comment.
