Anthropic's newly released Claude Fable 5 and its companion model Mythos 5 were taken offline within 72 hours of launch following pressure from both the White House and Amazon, according to multiple reports. The US government cited national security concerns, with officials specifically pointing to jailbreak risks — the ability of bad actors to manipulate the AI into producing dangerous or harmful outputs.
According to IT Pro, Anthropic suspended the Fable and Mythos systems "for all users" after the government raised those concerns. WinBuzzer reported that a warning from Amazon reportedly preceded the cutoff, suggesting the company — a major investor and cloud partner of Anthropic — played a role in the decision.
Anthropic is fighting back. According to MSN, the company is disputing the reasons given for the restrictions, saying the evidence does not justify such a broad move.
The backlash from the security community has been pointed. According to Sam Sabin writing for Axios, prominent cybersecurity leaders — including CISOs, security researchers, and executives at Adobe, Zoom, and Sophos — have written to the White House urging it to lift the ban. Their core argument: cutting off access to these tools hurts the defenders who rely on them far more than it hinders attackers.
The episode is also drawing international attention. According to VOI.id, India has highlighted the incident as evidence of the risks of technology dependence on the US, while Chosun Ilbo reported a growing push by nations worldwide to assert AI sovereignty in response.
This marks the first time the US government has reportedly forced a major AI lab to pull a commercial model for all users — setting a precedent that could reshape who controls access to powerful AI, and on whose terms.