U.S. orders Anthropic to disable Fable 5 and Mythos 5 after Amazon flags potential cyberattack data leakage risk

The U.S. government ordered Anthropic to immediately disable its frontier AI models Fable 5 and Mythos 5 after Amazon identified a potential risk that the models could leak information relevant to cyberattacks. Anthropic, which has received about $8 billion in investment from Amazon, became a high-profile example of escalating regulatory intervention in advanced AI. The episode involved no financial figures or product sales, but signals a de facto pre-deployment review for commercial frontier models that could add policy uncertainty for big tech firms relying on large-model commercialization.