The U.S. government ordered Anthropic to block all foreign nationals from its Fable 5 and Mythos 5 AI models. This export-control directive, issued Friday evening, applies to foreign individuals both inside and outside the United States. The ban includes Anthropic’s own employees. Anthropic disabled the models globally just days after launch to ensure compliance.
Officials cited national security concerns but provided no specific details for the directive. Anthropic attributes the order to a discovered jailbreak method that bypasses Fable 5 safety features. The company describes the vulnerability as a narrow potential jailbreak for minor, known issues.
This action marks a significant escalation in federal oversight of the AI industry. The directive follows ongoing tensions between Anthropic and the administration regarding AI safety.