US Government Orders Anthropic to Halt Fable 5 and Mythos 5 Access Over Security Fears
Following a government directive related to national security risks, Anthropic has suddenly taken its most advanced AI models offline.
Following receipt of an export control directive from the US government that cited national security risks, Anthropic announced it has suspended access to its Fable 5 and Mythos 5 AI models.
According to a statement released on Friday by Anthropic, the company was served with the directive at 5:21 pm ET, ordering it to halt all access to Fable 5 and Mythos 5 for any foreign national, regardless of whether they are located within or outside US borders, including foreign nationals employed by Anthropic itself.
To guarantee full compliance, Anthropic immediately took the models offline for every user. The company clarified that its other AI models, including Opus 4.8, remain unaffected by this action.
We are complying with the government's legal directive and are removing access to Fable 5 and Mythos 5 for all users.
Anthropic
This directive arrives mere days following Anthropic's launch of Fable 5 and Mythos 5, a pair of advanced AI models developed on the foundation of Mythos Preview, a multi-purpose language model that the company had earlier stated discovered thousands of security vulnerabilities in essential software systems.
According to Anthropic, authorities have not disclosed specific information regarding the purported threat, though the company indicates it understands the government's concerns center around a potential "jailbreak" technique that could circumvent Fable 5's built-in safety measures.
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.
Anthropic
The company emphasized that a non-universal jailbreak represents a significantly lesser risk compared to a "universal jailbreak," which would enable widespread circumvention of a model's protective safeguards.
We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people. If this standard was applied across the industry, we believe it would essentially halt all new model deployments for all frontier model providers.
Anthropic
The company stated its belief that the government's order stems from a misinterpretation of the situation and confirmed it is actively working toward reinstating user access at the earliest opportunity.