A legal tech startup’s Legion lawsuit over Anthropic models has become the first court challenge to a US government directive that stripped many users of access to two of the AI lab’s frontier systems. San Jose-based Legion filed suit in Washington, D.C., on Tuesday, asking a federal court to vacate the Commerce Department order and issue a preliminary injunction blocking its enforcement, according to Let’s Data Science, citing Reuters and Bloomberg.
The order at issue was issued on 12 June by the Commerce Department’s Bureau of Industry and Security. It directed Anthropic to bar foreign nationals, including its own foreign-national employees, from using Claude Fable 5 and Claude Mythos 5. Within hours of receiving that letter, and because it had no reliable way to sort users by nationality, Anthropic pulled both models for everyone. Lawfare described the practical effect as a global kill switch.
The Legion Lawsuit and What It Argues
Legion, founded in 2024, builds AI-powered litigation drafting software that helps lawyers automate pleadings, discovery requests, and other court documents. Its tools are built on top of frontier models, including those from Anthropic. The company employs Canadian nationals who work remotely from Canada, and it is those staff members whose access the directive effectively blocked.
In its complaint, Legion argued that it held a contractual right to use Fable 5. ‘Legion is a commercial customer of Anthropic with a contractual right and license to access and use the Fable 5 model, which was integral to building and operating its platform,’ the company wrote. ‘When the directive took effect, Legion lost the latest tool at the center of its development instantaneously.’
The company added that the order caused ‘immediate, irreparable, and existential’ harm, citing the rapid pace of AI development as the reason a competitive disadvantage could not be remedied later. The suit may be the first of several from Anthropic customers who lost access to the company’s most powerful models.
Anthropic is not a named party to the case. The company referred to a prior statement saying it was grateful to the administration for their ongoing partnership in working to get the matter resolved as quickly as possible, according to Let’s Data Science.
What Triggered the Export Control Order
The Commerce Department invoked what TechCrunch described as an ‘obscure export control directive’, a unilateral government action that did not appear to require court approval. The stated concern was the risk of a jailbreak of Fable 5.
Two separate lines of reporting point to what may have informed that concern. Forbes, citing the Wall Street Journal and Semafor, reported that the jailbreak concern has been linked to Amazon, with chief executive Andy Jassy allegedly telling Treasury Secretary Scott Bessent and other government officials that Amazon researchers used Fable 5 to obtain information that could be used in cyberattacks.
Separately, SiliconANGLE reported that the jailbreak concern is believed to relate to a statement from the Red Team lead at the UK’s AI Safety Institute, who said four days before the export ban that its cybersecurity team had made ‘substantial progress’ toward a universal jailbreak of Fable 5.
The National Law Review characterised the directive as an escalation in the use of export controls to restrict access to frontier AI models, noting that it applied to all foreign nationals ‘whether inside or outside the United States, including foreign national Anthropic employees.’
Anthropic’s Position and the Access Restoration
Anthropic itself pushed back on the scope of the restriction. When it first paused access, the company argued that cutting off all users over a ‘narrow potential jailbreak’ was excessive and would halt new model deployments entirely.
Last week, Anthropic restored access to Claude Fable 5 with nationality-based access controls and enhanced onboarding compliance screening in place. Claude Mythos 5 remains unavailable under the terms of the directive. The Commerce Department did not respond to a request for comment from Business Insider.
The Legion lawsuit introduces a new legal front into what has been, until now, a dispute managed through lobbying and negotiation. If the court grants even a preliminary injunction, it would set a precedent for how far the executive branch can reach in restricting access to commercial AI models through export control mechanisms, without congressional or judicial authorisation. The case’s next procedural step, the government’s response to Legion’s injunction request, will be the first real signal of how the administration intends to defend the directive in court.