Legion LegalTech Sues US Over Export Controls on Anthropic AI Models

Legion LegalTech has initiated a lawsuit against the US government over a directive from the Commerce Department that prohibits foreign nationals from accessing Anthropic's advanced AI models, specifically Fable 5 and Mythos 5. The lawsuit was filed on Tuesday in federal court in Washington, contesting a directive issued on June 12 by the Bureau of Industry and Security.

This directive mandated that Anthropic suspend access to its AI models for any foreign national, including individuals working within the United States. Consequently, Anthropic disabled access to these models for all customers globally on the same day, as it could not ensure compliance while maintaining broader access.

Legion, which is based in the US but employs Canadian software developers, utilizes Anthropic's models to create drafting and case management tools for attorneys. The company stated that the shutdown has immediately removed a critical model from its development efforts and has sidelined its engineering team. Legion characterized the impact as immediate, irreparable, and potentially existential, arguing that the fast-paced nature of the AI industry makes recovery from this setback difficult.

In its lawsuit, Legion is requesting the court to overturn the government directive and is also seeking a preliminary order to block its enforcement while the case is pending. The lawsuit names Commerce Secretary Howard Lutnick and other government officials as defendants, while Anthropic is not involved in the case. The Commerce Department and the White House have not yet commented on the lawsuit.

Anthropic has expressed appreciation for the administration's attempts to find a resolution and remains committed to protecting critical infrastructure and maintaining US leadership in artificial intelligence. The restrictions were implemented following national security concerns regarding the potential use of Fable 5 and Mythos 5 by foreign military or intelligence organizations. Anthropic has noted that the directive lacked specific details about the alleged threats and believes the government acted after identifying a technique that bypassed certain safeguards and a few known software vulnerabilities. The company contends that similar capabilities are already accessible through other public AI models, arguing that the reported bypass did not warrant the removal of the models from all customers. This legal dispute adds to ongoing tensions between Anthropic and the Trump administration, pa