Regulatory  ·  2026-06-21

US Commerce Department EAR Directive: Anthropic Fable 5 & Mythos 5 Foreign-National Access Suspended

RegulatoryHigh impactUnited States
On June 12, 2026 at 5:21 PM ET, Commerce Secretary Howard Lutnick invoked the Export Administration Regulations (EAR) and the Export Control Reform Act of 2018 (ECRA) to issue a binding directive to Anthropic, ordering the company to immediately suspend all access to its Claude Fable 5 and Claude Mythos 5 models by any foreign national — whether inside or outside the United States, including Anthropic's own foreign-national employees. Anthropic complied within hours by disabling both models globally for all customers (it could not practically distinguish foreign nationals from domestic users in real time). The stated government concern was a jailbreak that could allow access to Mythos 5's advanced cybersecurity capabilities. Anthropic disputed the severity. The directive remained formally in force as of June 19-20, despite President Trump telling Axios on June 19 he no longer views Anthropic as a national security threat — a political signal without legal effect. A bipartisan House letter (Reps. Liccardo, Obernolte, Lieu, Franklin) sent June 18 demanded Commerce explain the statutory authority invoked and criteria for restoring access, with a response deadline of June 26. This is the first-ever US government shutdown of commercial AI models under export control law.
This action establishes a watershed precedent: for the first time, the US government used export control authority — historically applied to physical goods and software code — to restrict foreign-national *access* to a commercially deployed AI model API. Every AI company deploying frontier models internationally must now treat export control compliance as a live operational risk. The directive demonstrates the government can compel a lab to take a model offline with 90 minutes' notice and no prior court approval. It also raises unresolved legal questions about whether EAR/ECRA authority extends to API usage, which Congress is now actively probing.
Audit all frontier AI model deployments for foreign-national access exposure. Build multi-provider fallback architectures. Review customer contracts for force majeure and government-action clauses. Map which users are foreign nationals if deploying high-capability models. Monitor Commerce Department response to the June 26 congressional deadline. Track whether the June 12 directive is formally rescinded or modified.
Sources
Anthropic Official Statement — Fable 5 & Mythos 5 Access SuspensionBloomberg — Lutnick's Anthropic Crackdown Claims New Power Over AI Models (June 19, 2026)Rep. Liccardo Press Release — Bipartisan Members Seek Transparency on Frontier AI Export Controls (June 18, 2026)Cryptopolitan — Trump Says Anthropic Is No Longer a National Security Threat (June 19, 2026)Forbes — Anthropic Disabled Fable 5 and Mythos 5 After US Export-Control Order (June 16, 2026)Volkov Law Blog — When the Government Pulls the Plug: Anthropic, Export Controls (June 2026)
See this in the live feed Explore related AI security and governance findings — updated every morning.
Open the feed →