A U.S. national-security directive suspends two frontier AI models
Anthropic disclosed that the U.S. government, citing national-security authorities, issued an export-control directive requiring it to suspend all access to two frontier models, Fable 5 and Mythos 5, for every customer. The directive followed a reported method for bypassing the models' safeguards. Anthropic said its own review found only minor, previously known vulnerabilities, that it disagrees with withdrawing a deployed model over a narrow finding, and that it is complying with the directive while working to restore access.
It is one more signal that frontier AI is now treated as national-security-grade by its developers and by government alike - the same fast-moving risk environment the NES Certified framework is built to evaluate at the level of a single home.