The Trump administration has ordered Anthropic to revoke access to its advanced Claude Mythos and Fable 5 AI models for all foreign nationals following national security and vulnerability concerns.

Key Points

• The White House mandated that Anthropic restrict access to its most powerful AI models to US citizens only, citing potential security risks.
• Anthropic chose to disable access to Claude Mythos and Fable 5 entirely rather than implement complex nationality-based screening processes.
• US officials expressed concern over South Korean telecom giant SK Telecom’s historical business ties to China and its access to the Project Glasswing program.
• Amazon researchers identified vulnerabilities in the Fable 5 model that could potentially allow users to bypass safety guardrails and access cyber-offensive capabilities.
• SK Telecom, which invested $100 million in Anthropic in 2023, denies having current ties to China that would compromise national security.

Why it Matters

This dispute highlights the growing tension between AI developers and the US government regarding the export and accessibility of powerful, dual-use artificial intelligence technologies. The decision sets a significant precedent for how AI labs must manage international partnerships and security protocols to comply with evolving federal oversight.