Amazon’s warning spurred U.S. shutdown of Anthropic’s top AI models
Amazon CEO Andy Jassy told senior Trump officials that researchers had cracked Anthropic’s Fable 5 model, exposing a jailbreak that could reveal software vulnerabilities. The warning triggered a U.S. national‑security order forcing Anthropic to suspend global access to its flagship models. The move underscores Amazon’s sway over AI policy and raises questions about export controls on advanced generative AI.