U.S. officials moved quickly to restrict access to a frontier AI model after a reported cybersecurity bypass during internal testing. The episode centers on Amazon researchers stress-testing Anthropic’s “Fable 5,” a “safe” variant of the company’s “Mythos” model. Amazon detected a jailbreak that could be used for cyberattacks and notified Anthropic. A key escalation followed when Amazon CEO Andy Jassy spoke with Treasury Secretary Scott Bessent after flagging concerns about cyber capabilities across frontier models. Soon after, the U.S. Department of Commerce imposed unprecedented export controls on Anthropic, prompting the lab to pull its most powerful models within days. The sequence signals how university and enterprise AI policies may be affected indirectly by export-control compliance requirements—particularly as educational research and labs increasingly depend on frontier AI access for coding, security, and automation work.