A U.S. legal technology company has filed suit against the federal government over a Commerce Department directive that forced AI developer Anthropic to shut off access to its Fable 5 and Mythos 5 models for all users worldwide, cutting off paying customers in the process and triggering what the plaintiff calls an existential threat to its business.
At a Glance
- Legion LegalTech Corp filed suit in Washington, D.C., federal court challenging a June 12 Bureau of Industry and Security order.
- The BIS directive required Anthropic to disable its Fable 5 and Mythos 5 models for any foreign national; Anthropic shut access for all customers to ensure compliance.
- Legion, headquartered in San Jose, California, builds drafting and case management software for attorneys and relies on Anthropic's models as core infrastructure.
- The company's Canada-based development team lost access immediately, and Legion argues the disruption is irreparable given the pace of frontier AI competition.
- Anthropic itself is already in separate legal disputes with the Trump administration in both Washington and California federal courts.
What the BIS Order Actually Did
On June 12, the U.S. Commerce Department's Bureau of Industry and Security issued a directive requiring Anthropic to disable its two most advanced commercially available models, Fable 5 and Mythos 5, for any foreign national. The scope of that phrase matters. Rather than attempt to screen each user against nationality, Anthropic pulled access for its entire global customer base the same day the order landed. That blunt compliance measure was, from the company's perspective, the path of least legal risk; from a customer's perspective, it was an immediate blackout with no warning.
Bureau of Industry and Security directives derive authority from the Export Administration Regulations and the Export Control Reform Act of 2018. Those statutes give the Commerce Department broad authority to restrict the export of items, including software and technology, deemed sensitive to national security. AI models can fall within that framework if the agency determines they constitute technology subject to export controls. The legal question Legion is now pressing is whether the June 12 directive was issued within those statutory bounds and whether it violated the Administrative Procedure Act, which requires agency actions to be neither arbitrary nor capricious.
Legion's Standing and the Harm Argument
Legion LegalTech Corp's standing to sue rests on a straightforward injury claim. The company is not Anthropic, and it had no say in how the directive was written or how Anthropic chose to comply. Yet its Canada-based software development team, the engineers building its attorney-facing drafting and case management platform, lost access to the core models the same day. In administrative law, third-party plaintiffs who suffer concrete, traceable harm from an agency action generally have standing to challenge that action under the APA, even if the agency's direct target is a different entity.
The lawsuit characterizes Legion's injury in notably strong terms: "immediate, irreparable, and existential." That framing is deliberate. To win a preliminary injunction, a plaintiff must typically satisfy a four-part test: likelihood of success on the merits, likelihood of irreparable harm absent relief, a balance of equities favoring the movant, and consistency with the public interest. Courts have held that business harm can qualify as irreparable when monetary damages are inadequate to make a party whole, and Legion is arguing exactly that. The company's lawsuit states that "competitive ground lost during a suspension cannot be regained after the fact," a claim aimed squarely at the irreparability prong.
Legion has also signaled it will seek a preliminary injunction to block enforcement of the directive while the case proceeds. That motion, if filed quickly, would force a hearing in the near term and could produce a temporary restraining order before any merits ruling.
Anthropic's Position and Its Own Litigation
Anthropic is not a named party in Legion's case, which is a meaningful distinction. The AI company referred inquirers on Tuesday to a prior statement expressing gratitude for what it described as the administration's "ongoing partnership in working to get this matter resolved as quickly as possible." That language suggests Anthropic is pursuing a negotiated resolution rather than an adversarial posture, at least on the BIS access question.
That said, Anthropic is simultaneously locked in separate litigation with the Trump administration in both Washington and California federal courts. Those cases stem from the government's reported effort to place Anthropic on a supply-chain blacklist after the company refused to allow military use of its AI models for domestic surveillance or fully autonomous weapons systems. The two sets of disputes are legally distinct, but they sit in the same political context: the administration pressing AI companies to cooperate with national security objectives, and at least some of those companies pushing back through the courts.
Because Anthropic is not a party to the Legion lawsuit, the AI company's own litigation posture will not directly bind or protect Legion. The San Jose company must win or settle on its own terms.
The Export Control Framework and AI Models
Export control law was originally designed for physical goods and hardware, but the Export Control Reform Act and subsequent regulations have extended its reach to software and technology. The question of whether a large language model constitutes an "item" subject to export licensing requirements has not been definitively resolved by federal courts, which is part of what makes the Legion case legally significant beyond the immediate business dispute.
If a court agrees that the BIS directive exceeded statutory authority, or that Anthropic's compliance response swept more broadly than the directive required, the ruling could establish important precedent for how export controls apply to AI model access. Conversely, if the government prevails, it would signal that agencies can restrict access to AI models on national security grounds with the same tools historically used to control hardware exports. Rules and outcomes vary by jurisdiction; what applies in a D.C. federal court may differ from how courts in other circuits analyze similar questions.
Frequently Asked Questions
Why did Anthropic shut off access for all customers rather than just foreign nationals?
The BIS directive required Anthropic to disable its Fable 5 and Mythos 5 models for any foreign national. Reliably verifying nationality for a global user base in real time presents significant compliance risk, so Anthropic appears to have made a conservative compliance decision by suspending access broadly. The company has not publicly detailed its internal reasoning.
What law gives the Commerce Department authority to restrict AI model access?
The Bureau of Industry and Security enforces the Export Administration Regulations under the authority of the Export Control Reform Act of 2018. Those rules govern exports of items, including technology and software, with potential national security implications. Whether an AI model's availability through an API constitutes an "export" under those rules is a legal question at the center of cases like this one.
What is the Administrative Procedure Act and why does it matter here?
The APA governs how federal agencies create and enforce rules. It allows courts to set aside agency actions that are arbitrary, capricious, not in accordance with law, or taken without proper procedure. Legion is invoking the APA to argue the June 12 directive should be vacated because it exceeded the agency's authority or was otherwise unlawful.
Does this ruling set a precedent for other AI companies?
A trial court ruling in this case would be persuasive but not binding on courts in other circuits. However, given the limited case law on export controls applied to AI models, any substantive ruling on the merits would likely be closely watched by the broader industry and by regulators.
What Comes Next
The Commerce Department and the White House had not responded to comment requests as of the filing. Legion's next procedural step is likely a motion for a preliminary injunction or temporary restraining order, which would put the case before a judge quickly. Anthropic's quiet diplomacy with the administration could also produce a resolution that moots the litigation before any ruling, though that outcome depends entirely on negotiations that are not public. The broader question of how export control law applies to frontier AI models now has at least one active federal case testing its limits.
This article is general informational content about legal and regulatory developments and does not constitute legal advice. Laws and regulations vary by jurisdiction, and anyone affected by similar issues should consult a qualified attorney.
