House Lawmakers Press Lutnick on Legal Basis for Anthropic AI Ban

When the Commerce Department's Bureau of Industry and Security told Anthropic in mid-June to cut off foreign nationals from its Fable 5 and Mythos 5 models, it acted not through a court order or an act of Congress but through a letter from Howard Lutnick. According to The Washington Post, a bipartisan group of House lawmakers has since sent a letter to Lutnick pressing him on exactly what statutory authority he invoked when ordering Anthropic to block foreign nationals from accessing an AI API rather than controlling physical goods or exportable code.

That distinction matters more than it might look. Export control law, specifically EAR § 744.22, gives the Bureau of Industry and Security authority to restrict the "export of items." Legal analysts have argued the framework was not written with API inference access in mind, and that there may simply be no "export" here under the law's existing definitions. If Lutnick cannot supply a clean statutory answer to Congress, the directive looks like an improvised use of peacetime trade law applied to something it was never designed to cover.

The administration's stated rationale, according to Axios, is that an unnamed company claimed it had jailbroken Mythos, which alarmed officials about potential national security risks. Because Anthropic could not reliably screen users by nationality, it disabled both Fable 5 and Mythos 5 entirely rather than attempt partial restriction. The House lawmakers specifically flagged that the action "may have opened a precedent with significant implications for the AI industry."

What the reporting does not give you: the identity of the company that allegedly jailbroke Mythos, the nature of the capability demonstrated, or whether Anthropic is considering a legal challenge to the directive. Those gaps matter because the jailbreak claim is the entire factual predicate for the national security argument, and it has not been independently verified or publicly detailed.

The bigger structural question is whether a Commerce Secretary can unilaterally suspend access to any AI model deemed sensitive, via a letter citing a statute written for physical goods, with no transparent review process. Congressional oversight is the appropriate check, but oversight letters without enforcement teeth rarely move fast enough to matter on a live product decision. Anthropic's users outside the US are, in the meantime, simply offline.