US government orders Anthropic to shut down most powerful AI models
Original: Anthropic’s safety warnings may have just backfired — the government has pulled the plug on its most powerful AI
Why This Matters
First major government shutdown of commercial AI models raises questions about AI regulation
The US government ordered Anthropic to immediately disable Claude Fable 5 and Claude Mythos 5 models worldwide on Friday, citing national security concerns. The directive affects all users globally, not just foreign nationals targeted by export controls.
Anthropic received the government directive at 5:21 PM ET on Friday to shut down access to its two most advanced AI models. Mythos 5, previewed in April, was kept restricted due to exceptional ability to find software security vulnerabilities across major operating systems and browsers. It was shared with only 50 vetted organizations including Amazon, Apple, Google, Microsoft, and CrowdStrike through Project Glasswing for defensive cybersecurity work. Fable 5, released three days prior, was a commercially available version with guardrails blocking high-risk responses in cybersecurity and biology. According to Vals AI benchmarks, it was the most capable publicly available AI model. The government cited a claimed jailbreak of Fable 5, though Anthropic says only verbal evidence of a 'potential narrow, non-universal jailbreak' was provided. Anthropic argues similar capabilities exist in other public models like OpenAI's GPT-5.5 and are routinely used by cybersecurity professionals.