Anthropic, the California‑based AI firm, pulled its latest model, Claude Fable 5, from service after U.S. authorities issued a security warning.
The company said, in a statement on its website, that it was ordered to suspend foreign nationals from using Fable 5 and Mythos 5, calling the tools “too powerful” and noting concerns over a potential jailbreak technique that could bypass safety measures.
Officials did not name a specific vulnerability, but the U.S. Department of Commerce reportedly cited the possibility of a jailbreak that “could expose a small number of known minor flaws”. The disruption was significant given Claude Fable 5 was publicly released just weeks after a private beta that revealed similar weaknesses.
Anthropic’s halt comes amid an ongoing lawsuit between the company and the Trump administration, which has already designated the firm a “supply‑chain risk”—an unprecedented move against a U.S. technology company.
The court’s decision that the Pentagon’s directive could not be enforced means some government agencies can continue using Anthropic’s AI for the time being, but the controversy underscores how quickly policy can shift in response to perceived AI threats.
Experts say the incident highlights growing scrutiny as AI systems grow more capable, raising questions about how to balance rapid innovation with national security.
