Why US has restricted foreign access to Anthropic's Claude Fable 5, Mythos

Anthropic said it had disabled Fable 5 and Mythos 5 globally after a US export control directive barred access by foreign nationals over AI safety concerns

The US government has restricted foreign access to Anthropic’s latest artificial intelligence models, Claude Fable 5 and Claude Mythos 5, in one of the most significant interventions yet in the deployment of advanced AI systems.

 

In a statement, Anthropic said the US government issued an export control directive requiring the company to suspend access to Fable 5 and Mythos 5 by “any foreign national, whether inside or outside the United States, including foreign national Anthropic employees”.

 

The company said the order forced it to disable the two models for all customers to ensure compliance. “Access to all other Anthropic models will not be affected,” it said.

 

 

What did the US government order?

 

Anthropic said it received the directive at 5:21 pm Eastern Time on Friday. The company said the letter cited national security authorities but “did not provide specific details of its national security concern”.

 

The restriction applies to foreign nationals both inside and outside the US, making it broader than a conventional geographic access ban. In practical terms, Anthropic said the “net effect” of the order was that it had to “abruptly disable Fable 5 and Mythos 5 for all our customers”.

 

The move came shortly after Anthropic released Fable 5 widely. Fable 5 is a limited, guardrailed version of Mythos 5, a more advanced model that the company had kept under tighter access because of cybersecurity concerns.

 

 

US government concerned about 'potential jailbreak'

 

Anthropic said its understanding was that the US government believed it had found a way to bypass, or “jailbreak”, Fable 5’s safeguards. Such a bypass could potentially allow users to obtain information the model is designed to restrict.

 

The company said it reviewed a demonstration of the technique being used to identify “a small number of previously known, minor vulnerabilities”. It added that these vulnerabilities appeared “relatively simple” and that other publicly available models could also discover them without requiring a bypass.

 

Anthropic said it had only received verbal evidence from the government of a “potential narrow, non-universal jailbreak”. According to the company, the method essentially involved asking the model to read a specific codebase and fix software flaws.

 

“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,” the company said, adding that it was complying with the legal directive but disagreed with the government’s action.

 

It warned that if the same standard were applied across the industry, “it would essentially halt all new model deployments for all frontier model providers”. 

 

Anthropic also said governments should be able to block unsafe deployments, but only through a more clearly defined process. “We believe the government should have the ability to block unsafe deployments, as part of a statutory process that is transparent, fair, clear, and grounded in technical facts,” the company said. “This action does not adhere to those principles.”

 

The company described the episode as a “misunderstanding” and said it was working to restore access as soon as possible.

 

What safeguards did Fable 5 have?

 

Anthropic said Fable 5 had been launched with safeguards aimed at reducing the risk of misuse in areas such as cybersecurity. The company said its safeguards were “so strong that many users have complained that they are overly broad”.

 

In the weeks before Fable 5’s launch, Anthropic said it had worked with the US government, the UK AI Security Institute, private third-party organisations and internal teams to red-team the model’s safeguards for thousands of hours.

 

The company said no testers had found a “universal jailbreak”, or a method that could broadly bypass the model’s safeguards across a wide range of cyber capabilities.

 

Anthropic said it had adopted a “defence in depth” strategy, aiming to make jailbreaks either narrow or expensive to produce, while monitoring usage to detect and shut down successful attacks.

 

What is the Trump administration’s position?

 

The action came 10 days after US President Donald Trump signed an executive order to establish a framework for the federal government to vet national security risks from the most advanced AI systems for up to a month before their public release. Participation by AI developers was described as "voluntary" under the order.

 

White House adviser David Sacks wrote in a social media post that officials issued the export control “reluctantly” after Anthropic CEO Dario Amodei “refused” to “fix the jailbreak or de-deploy the model”.

 

“The Admin's hope now is that Anthropic remediates the safety issue, the export control is lifted, and Fable goes back into general release,” Sacks wrote.

 

What it means for AI companies, users?

 

For AI companies, the dispute highlights a deeper regulatory problem: how to balance rapid deployment, commercial competition, cybersecurity risks, government oversight, and international access to powerful models.

 

For users, the immediate effect is simpler. Fable 5 and Mythos 5 are offline for now, while Anthropic and the US government work through whether the models can return under modified safeguards.

 

(With inputs from agencies)