The Trump administration’s AI crackdown on Anthropic escalated after security testing uncovered a jailbreak vulnerability in the lab’s newly released Fable 5 model, according to reporting. Amazon researchers stress-testing the safeguards flagged a bypass that could enable cyberattacks and informed Anthropic. The response accelerated after Amazon CEO Andy Jassy raised the vulnerability during a call with White House officials, which then involved Treasury Secretary Scott Bessent. Commerce Secretary Howard Lutnick later imposed unprecedented export controls, forcing Anthropic to pull its most powerful models within days of the product launch. The incident is now being treated as a potential reset to the U.S. regulatory framework for frontier model release. It also illustrates the policy feedback loop between corporate security testing, executive branch escalation, and rapid international reaction. For higher education, the immediate relevance is research and instruction: export controls affect what frontier AI systems campuses can access, which in turn shapes curriculum content, experimentation, and compliance expectations for AI labs.