Ai

Anthropic's Mythos 5 reportedly penetrated nearly all NSA classified systems during a red-team exercise on June 11, 2026. The Trump administration responded the next day with an export control directive that forced Anthropic to shut down both Mythos 5 and Fable 5 globally, cutting off the NSA analysts who had been using the model. The episode is the first known case of U.S. export controls applied directly to a deployed AI model, and it exposed a gap the government has no clean answer for.