A U.S. national security and export-control response disrupted access to Anthropic’s newest AI models after an Amazon-discovered vulnerability scenario. Reported details cite an exchange in which Amazon cybersecurity researchers prompted the models with a “fix this code” instruction to generate patches, raising concerns the approach could be used offensively. As a result, the Commerce Department imposed export controls that disabled Fable 5 and Mythos 5 for foreign nationals, including non-citizen employees working inside the U.S.