The Trump administration has ordered Anthropic, one of the United States' leading artificial intelligence companies, to disable its two most advanced AI models — Fable 5 and Mythos 5 — after Amazon researchers identified a way to bypass the models' safety guardrails, raising concerns about potential cyberattack and bioweapons risks. The White House moved rapidly to impose foreign export restrictions, leaving Anthropic with no option but to cut off access for all users. The unprecedented intervention signals a sharp shift in the administration's approach to AI governance, with industry figures and cybersecurity experts warning that the action removes powerful tools from legitimate researchers while doing little to address vulnerabilities present in other publicly available models.