US Oversight Restricts Access to Frontier AI Models - Let's Data Science

On June 12, the U.S. Commerce Department ordered restrictions on access by foreign nationals to Anthropic's Claude Fable 5 and Mythos 5 AI models. In response, Anthropic disabled both models for all users, citing that real-time nationality checks were not feasible. The restrictions were lifted on June 30 after the company implemented safeguards and underwent a government review.

The temporary ban and subsequent restoration highlight ongoing tensions between AI safety concerns and accessibility. Anthropic had previously restricted Mythos due to safety issues, later clearing it for use after working with U.S. intelligence agencies. Claude Fable 5 is a newer frontier model. The Commerce Department order reflects government efforts to prevent foreign adversaries from exploiting advanced AI, though the abrupt global shutdown illustrates challenges in enforcing such controls.

Anthropic has not detailed the specific safeguards added during the 18-day pause, but stated they satisfied government requirements. The episode underscores the difficulty of balancing national security with the open development of AI, and the practical obstacles to nationality-based access restrictions. No further regulatory actions have been announced.