On June 23, 2026, Legion LegalTech Corp filed suit against the US government, becoming the first known legal challenge to export-control authority targeting frontier AI models. The case stems from a Bureau of Industry and Security (BIS) directive issued June 12, ordering Anthropic to block foreign nationals from accessing its advanced models, Fable 5 and Mythos 5. Anthropic complied immediately, disabling both models globally-not just for flagged users-which disrupted Legion’s Canadian development team that relies on them for legal drafting and case management software.
The BIS framed the directive as a national security measure, citing jailbreak vulnerabilities exploitable by foreign actors. Legion argues the order caused unlawful disruptions and represents an overreach of export controls traditionally focused on hardware and dual-use technologies. The lawsuit raises a fundamental question: whether the US government can use export-control powers to restrict access to commercial AI tools that are used for mundane tasks.
The legal fight has rippled into markets for decentralized AI. Tokens tied to such projects rose on the logic that centralized AI platforms, which can be switched off overnight by regulatory fiat, create operational risk. This case will help define how companies build and rely on AI services going forward.