UNITED STATES PERSPECTIVE

Washington has just crossed a line no one had dared cross: treating an AI model as an export-controlled weapon. On Friday June 12 at 5:21 p.m. ET, Commerce Secretary Howard Lutnick sent a letter to Anthropic CEO Dario Amodei ordering it to suspend all access to Mythos 5 and Fable 5 'by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.' Unable to sort users by nationality, the company had to 'abruptly disable' both models for all customers. The American press reconstructed the sequence: Amazon — itself a major Anthropic investor — called the administration on Thursday night to flag that it had 'jailbroken' Mythos and reached portions of the model posing a national-security risk; at least five other companies relayed the alarm. An official told Axios the administration had tried to persuade Anthropic to delay the launch, to no avail, hence the export letter. The context is explosive: Anthropic already sits on a Pentagon blacklist — deemed too dangerous for the government's own use — after refusing to let the military deploy its models for domestic surveillance and fully autonomous weapons. The company is thus 'too dangerous for public use and too dangerous for state use' at once, months from a confidentially filed IPO. The American debate polarizes instantly: legitimate national security for some, an unprecedented move raising 'constitutional and commercial questions' about restricting software for others. Anthropic pleads 'misunderstanding' and disputes that a 'narrow potential jailbreak' warrants recalling a model deployed to hundreds of millions of users.