🙀 US Gov Shuts Down Claude Fable
The Neuron · Grant Harvey · 2026-06-14
The US government issued an emergency export-control directive forcing Anthropic to disable its new Claude Fable 5 and Mythos 5 models for all users after Amazon researchers flagged a jailbreak to federal officials, marking the first known government-mandated post-launch shutdown of a frontier AI model.
Appears in
Extraction
Topics: ai-governanceexport-controlsanthropicclaude-fableai-safety-policy
Claims
- The US government issued an export-control directive on June 12 blocking foreign nationals from accessing Claude Fable 5 and Mythos 5, citing national security.
- Because Anthropic cannot verify user nationality in real time, it disabled both models for all customers globally.
- Amazon researchers identified and escalated a Fable 5 jailbreak to US officials, directly triggering the government action.
- This event marks a shift from company-led AI safety decisions to government-mandated model shutdowns after public launch.
- The precedent could invite retaliatory restrictions by China that cut US companies off from frontier open-weight models, harming competitive AI pricing.
Key quotes
This moves the AI governance landscape from 'companies decide what to ship' to 'the government can pull a frontier model after launch now'.
Anthropic kinda asked for this? They've been asking for the government to regulate the industry and block unsafe deployments like this.
if a model can disappear overnight, procurement teams must now consider the geopolitics of a model's origin