Statement on the US government directive to suspend access to Fable 5 and Mythos 5

The United States government has issued an export control directive requiring Anthropic to suspend all access to its Fable 5 and Mythos 5 artificial intelligence models. This mandate applies to all foreign nationals, both inside and outside the United States, and has forced Anthropic to disable these specific models for all customers immediately. Access to other Anthropic models remains unaffected.

The government directive cites national security concerns regarding potential methods to bypass, or jailbreak, the safety protocols of Fable 5. Anthropic reports that the government identified a narrow, non-universal technique that allows the model to identify software flaws. Anthropic argues that this capability is already widely available in other publicly accessible models and is commonly used by cybersecurity professionals to defend systems.

Anthropic maintains that its defense-in-depth strategy for Fable 5 is robust and that its safeguards are more effective than those of previous models. The company states that it has not received evidence of any harmful results stemming from the identified jailbreak. While Anthropic is complying with the legal order, it disagrees with the government's decision, noting that applying such a standard would effectively halt all new AI model deployments across the industry. The company is currently working to resolve what it characterizes as a misunderstanding to restore access to the models.