“We received the directive from the government today at 5:21pm (ET). The letter did not provide specific details of its national security concern. Our understanding is that the government believes it has become aware of a method of bypassing, or ‘jailbreaking’ Fable 5,” Anthropic said in a June 12 blog post.

“We reviewed a demonstration of this specific technique being used to identify a small number of previously known, minor vulnerabilities. These vulnerabilities all appear relatively simple, and we have found that other publicly available models are able to discover them as well without requiring a bypass.”

The company insists, however, that its safeguards are strong and said it had worked with the government to red-team Fable’s safeguards in the lead up to its launch.

“These tests showed that Fable’s safeguards are substantially more effective than those of any previously deployed model,” Anthropic said.

“No testers have yet been able to find a universal jailbreak – a jailbreak method that can very broadly bypass the model’s safeguards, unblocking a wide range of cyber capabilities.”

Anthropic added that it understands that only one potential jailbreak was shared with the government, although it has only been provided with “verbal evidence” that it took place.

“As we have stated publicly, 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. This action does not adhere to those principles,” Anthropic said.

“We apologise for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.”

The suspension comes just one week after the general launch of Fable 5 and the expansion of access to its Mythos 5 model.

Cynthia Lee, APAC Vice President at software firm Delinea, said the US government’s actions were “likely to reverberate in Asia Pacific”.

“One way these new regulatory steps may impact APAC companies is by increasing scrutiny of whether their services and products sold into the US use non-American AI models, in line with legislation approved in 2025,” Lee told Cyber Daily.

“The shift in the US also has the potential to influence how regional policymakers and industry think about regulation in their own countries. The likely direction is towards incentive frameworks that encourage AI technology vendors to weigh the societal implications of the solutions they bring to market, rather than broad regulatory regimes. In that approach, the responsible providers tend to become the examples others follow."