Anthropic has disabled all customers' access to its Fable 5 and Mythos 5 AI models after receiving a government directive on Friday, June 12, citing national security concerns, according to Engadget. The company stated that the US government wanted it to suspend all foreign nationals' access to the newly launched models, whether they are inside or outside the US, including Anthropic employees. All other models and the Claude chatbot are not affected.
The Government Directive
Anthropic received an order from the US government on June 12 requiring it to block foreign nationals' access to Fable 5 and Mythos 5, according to the company's announcement. The government did not specify the exact national security concerns, but Anthropic believes it is because officials heard about a method for jailbreaking Fable 5. The company has been vocal about AI safety and previously warned about the need for regulation, but it disagrees that a potential jailbreak should be cause for recalling a commercial model. Engadget reported that the company stated: "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."
Which Models Are Affected?
| Model | Availability | Status |
|---|---|---|
| Fable 5 | Public (launched June 9) | Access blocked for all customers |
| Mythos 5 | Project Glasswing partners only | Access blocked for all customers |
| Claude chatbot | Public | Unaffected |
| Other Anthropic models | Various | Unaffected |
Fable 5 was launched on June 9 and designed to bring many of Mythos' capabilities to the public. Mythos is Anthropic's state-of-the-art cybersecurity model, previously only available to Project Glasswing partners. According to Engadget, Fable's capabilities "exceed" any previous Anthropic model. During tests, Fable beat the game Pokémon FireRed, while Claude failed to beat Pokémon Red.
The Jailbreak Concern
The government provided verbal evidence of one potential narrow, non-universal jailbreak that an unnamed entity shared with officials. Anthropic implemented strong safeguards to "reduce the likelihood that Fable is misused for tasks related to cybersecurity" and noted that many users complained safeguards were "overly broad." The company acknowledged that no provider can ensure perfect resistance to jailbreak attempts and that every model is vulnerable to jailbreaks made especially for it. Its defense strategy aimed to "make jailbreaks either narrow (in the case of non-universal jailbreaks) or very expensive to produce (in the case of universal jailbreaks), and to combine this with thorough monitoring to quickly detect and shut down any successful attacks."
Anthropic's Response and Arguments
Anthropic promised to share more details over the next 24 hours but clarified its disagreement with the government's action. The company has previously warned about the need for AI safety and believes any government intervention should follow a transparent statutory process. Engadget noted that Anthropic's announcement emphasized that the current directive does not meet those principles. For enterprise customers using Fable 5 or Mythos 5, the sudden block means immediate disruption to any workflows relying on these models. The company has not provided a timeline for restoring access.
Sources:
