The US Commerce Department ordered AI company Anthropic in June 2026 to immediately block all foreign nationals from accessing its two most powerful AI models — Claude Fable 5 and Mythos 5 — citing national security concerns, in what appears to be the first use of export control law to restrict access to an AI service. Anthropic (a leading US artificial intelligence company) complied within hours, ultimately disabling the models for all users after finding no quick way to verify nationalities globally; the restrictions were lifted only after the company added safety guardrails that significantly degraded the models' capabilities, with one test finding the updated Fable 5 completed just 3 of 12 tasks it could previously handle. Legal scholars question whether the order was lawful, noting that the Export Control Reform Act of 2018 was written with physical goods in mind, that the Commerce Department's own past guidance had excluded remote software access from export rules, and that the directive — issued unilaterally, in secret, with no expiry date and no allied coordination — bypassed procedural safeguards normally required for emergency technology controls.