US Orders Anthropic to Block Foreign Access to Claude Models Over National Security Concerns

Summary

The US government has ordered Anthropic to prevent all foreign nationals from accessing its advanced AI models Fable 5 and Mythos 5, citing national security concerns. According to a blog post published Friday, the ban applies to all foreign users globally as well as foreigners currently working at Anthropic and in the United States. The government's primary concern appears to be the models' cybersecurity capabilities, particularly Mythos 5's exceptional ability to detect software vulnerabilities that have remained hidden for decades—a capability that could potentially be weaponized if misused.

Anthroporic received the order at 5:21pm GMT on Friday with minimal explanation from authorities and was forced to implement the block immediately. The company said it reviewed a government report that appears to reference the models' limited capability to review and correct program code, and has expressed disagreement with the decision. Anthropic argues that competing models, such as OpenAI's GPT-5.5, possess similar capabilities and that Fable 5's safety measures have been extensively tested.

The order reflects escalating government scrutiny of advanced AI systems and their potential national security implications. Anthropic has been proactive in raising concerns about rapid AI development, recently proposing that leading AI companies coordinate to pause development of advanced systems to mitigate risks of losing human control over increasingly capable technologies.

• The company disputes the decision, noting that competing AI systems like OpenAI's GPT-5.5 have similar capabilities and Fable 5's safety measures have been rigorously tested