U.S. regulators force Anthropic to recall Claude after jailbreak audit
Federal authorities have ordered Anthropic to pull its flagship Claude model from commercial use following a safety audit that identified a jailbreak vulnerability—the first government-mandated recall of a widely deployed commercial AI.
Anthropic's flagship Claude model is no longer available to its hundreds of millions of users after U.S. government regulators ordered the company to recall the AI following a safety audit. The recall marks the first time federal authorities have pulled a widely deployed commercial AI model from the market. Anthropic disclosed the order in a blog post this week, writing that the company "disagrees that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people."
The move appears to stem from Anthropic's own safety reporting practices. The company has been vocal about AI risks and has published detailed safety warnings alongside its model releases—a transparency approach that may have provided regulators the evidence needed to justify the recall. The specific jailbreak vulnerability has not been publicly detailed, but Anthropic characterized it as narrow in scope. Anthropic has positioned itself as the safety-conscious alternative in the large language model market, regularly publishing research on constitutional AI and model behavior. That same transparency may now be working against the company—detailed vulnerability reports meant to demonstrate responsible development have instead given regulators concrete grounds for intervention.
