U.S. Commerce Department Requires Anthropic to Suspend Fable 5 and Mythos AI Models for Foreign Users After Security Flaws Disclosed

The U.S. Commerce Department's Bureau of Industry and Security (BIS) issued an administrative order mid-June requiring Anthropic to suspend API access to its flagship model Fable 5 and cybersecurity-focused model Mythos for foreign nationals—including overseas customers and non-U.S. staff. Anthropic was given only 90 minutes to comply and, unable to instantly verify user nationalities globally, shut down both models worldwide, affecting even Five Eyes allies and the UK AI Safety Institute. The action marks the first U.S. administrative ban on publicly deployed AI models, shifting regulation from chip restrictions to model access control.

The restriction was triggered by Amazon researchers discovering a potential jailbreak vulnerability in Fable 5 that bypasses security classifiers to expose network analysis capabilities. According to Senate Intelligence Committee Vice Chair Mark Warner, NSA officials stated Mythos penetrated classified system testing in mere hours during red team exercises. Anthropic plans to restore Fable 5 access to U.S. citizens by July 8 with identity verification mechanisms in place.