Amazon and five other tech companies reported security concerns about Anthropic's Fable model to the Trump administration, triggering a rapid government shutdown of the AI system within hours.

What happened: Amazon CEO Andy Jassy, joined by at least five other company executives, brought concerns about security risks in Anthropic's Fable model to senior Trump administration officials on Thursday evening and Friday morning. Amazon submitted a report claiming to show how parts of the model could be unlocked through jailbreaking. The White House called for the model to be pulled voluntarily; when Anthropic refused, the government issued an export control order at 5:20 p.m. ET, giving the company 90 minutes to comply. Anthropic shut the model down by 10:00 p.m.

Why it matters: Amazon is one of Anthropic's largest investors and runs the company's AI infrastructure, making this an unusual case of a major investor directly reporting its own portfolio company to the government. The speed and force of the government response—moving within hours and using export controls—signals an effort to establish what one source called a "de facto licensing regime" rather than respond to a genuine security threat. Security expert Katie Moussouris, who reviewed Amazon's report, called the government's response "wildly disproportionate," arguing that what Amazon flagged was a defensive technique, not a jailbreak.

What to watch: The White House's frustration with Anthropic appears rooted less in the specific security risk than in what officials saw as a "lack of seriousness" in how the company handled the release—Anthropic pushed Fable out before a security review process that the administration had requested was even in place. According to The Information, the export restriction is "unlikely" to be extended to other AI labs, suggesting this action targets Anthropic specifically.