For the first time in history, a government has forcibly shut down access to cutting-edge commercial AI models for millions of users around the world.
This week Anthropic released Fable 5 and Mythos 5 — its flagship AI systems. Mythos was built specifically for deep code vulnerability research and distributed only to security researchers; the base Fable went into public release.
Emergency directive
On Friday at 5:21 PM, Anthropic's leadership received an emergency directive from the US Department of Commerce. Officials cited a national security threat: apparently someone had found a way to jailbreak the models — bypassing built-in safety systems and content filters — and force Fable to hunt for critical vulnerabilities in third-party software.
Officials demanded the immediate cutoff of access for all foreign nationals, including Anthropic's own overseas employees. Filtering users by citizenship in the cloud turned out to be technically impossible on short notice, so the company was forced to take drastic measures.
What Anthropic did
- Shut down Fable 5 and Mythos 5 servers entirely for all customers worldwide, including paid subscribers.
- Left only older, less capable model generations available.
- Published an open letter calling the government's action a "misunderstanding" — officials hadn't provided any coherent technical details of the alleged vulnerability.
Market reaction
The reaction was swift: the tech community is already worried that this precedent could paralyze every major player and freeze deployment of any advanced technology. The industry has reached a dangerous threshold where regulators can pull the plug on a cutting-edge product with a single letter — no public evidence of actual harm required.