Judge Lin: Trump Anthropic Ban Illegal, Citing First Amendment Retaliation

• A California federal judge ruled that the Trump administration illegally retaliated against AI company Anthropic.
• The retaliation stemmed from Anthropic's refusal to allow its Claude chatbot for lethal autonomous warfare or mass surveillance.
• Judge Rita Lin found the government's "national security risk" justification was a pretext for punishing Anthropic's protected speech.
• The government's own admissions and contradictory actions, such as continued negotiations and deployment of other Anthropic AI, undermined its claims.
• The ruling reinforces First Amendment protections for companies in government contracting against retaliatory actions by the executive branch.

The empty invocation of national security is not a blank check to punish and retaliate against government critics.

A federal judge in California recently invalidated the Trump administration's prohibition on the military's use of artificial intelligence technology developed by Anthropic. U.S. District Judge Rita Lin determined that former President Donald Trump and then-Secretary of War Pete Hegseth engaged in illegal retaliation against the AI firm. This significant ruling came after Anthropic declined to remove specific restrictions on its Claude chatbot, which prevented its application for mass surveillance of American citizens and lethal autonomous warfare.

President Trump had publicly labeled Anthropic as "leftwing nut jobs" following the company's refusal to grant unrestricted access to its AI for military purposes. Secretary Hegseth subsequently issued a directive formally designating Anthropic as a "supply chain risk to national security," effectively banning any company working with the military from collaborating with Anthropic. Judge Lin, an appointee of President Joe Biden, explicitly rejected the government's stated justification for this designation, pointing to internal communications that revealed the ban's true intent was to publicly chastise Anthropic for its criticism of the administration.

In her decision, Judge Lin underscored the limitations of executive power, stating, "The empty invocation of national security is not a blank check to punish and retaliate against government critics." This declaration highlights the court's scrutiny of government actions when they appear to target protected speech, particularly in the context of government contracts and national security claims. The ruling reinforces First Amendment protections for companies engaging with the government, setting a precedent against using national security as a pretext for punitive measures.

The conflict between Anthropic and the Trump administration originated from protracted private negotiations concerning a new Pentagon contract. Anthropic CEO Dario Amodei engaged in discussions with Under Secretary of Defense for Research and Engineering Emil Michael, during which the Department of Defense pressed Anthropic to eliminate all usage restrictions on its Claude chatbot. While Anthropic agreed to relax most of its limitations, the company steadfastly refused to permit the AI's deployment for lethal autonomous warfare or widespread surveillance.

In February, Amodei publicly articulated Anthropic's ethical stance, stating that the company could not, in good conscience, provide such extensive access to its technology. The administration's response was swift and public. The very next afternoon, President Trump utilized Truth Social to issue an order directing all federal agencies to immediately cease using Anthropic’s technology. Concurrently, Secretary Hegseth posted his own directive on X, formalizing Anthropic's designation as a "supply chain risk" and prohibiting military contractors from any engagement with the company.

Central to the government's defense was its assertion that Anthropic posed a "supply chain risk to national security." This claim was supported by an internal memo drafted by Under Secretary Michael, which warned that Anthropic's AI could be remotely disabled or altered during active military operations. However, during the course of litigation, the government conceded that Anthropic lacks such "backdoor" access to its models once they are deployed. Furthermore, it admitted that the Claude chatbot presented no greater risk in this regard than any other "black box" AI model.

Judge Lin found these admissions significantly weakened the government's justification. In her assessment, the only factor distinguishing Anthropic from other vendors was an issue of "trust," rather than a genuine security vulnerability. Her ruling explicitly stated that Anthropic’s protected speech, which addressed a matter of considerable public importance, served as a substantial motivating factor for the defendants' actions. She concluded that the retaliatory measures would not have been taken had it not been for the administration's desire to make an example of Anthropic for its public stance on critical issues within the contracting dispute.

Further undermining the government's national security narrative were several contradictory actions. Days before the official ban, Secretary Hegseth had considered invoking the Defense Production Act, a measure that would have treated Anthropic as essential to national security rather than a threat. Moreover, Under Secretary Michael continued to negotiate contract language with Anthropic CEO Amodei even after the ban was in effect. An email from Michael to Amodei on March 4, the same day Anthropic received Hegseth's formal designation letter, indicated that they were "very close" to an agreement.

The government's continued engagement with Anthropic's newer Mythos model also raised questions. Multiple agencies, including the National Security Agency, began deploying the Mythos model around the same time officials were publicly branding Anthropic a security threat. The defendants provided no evidence to explain this discrepancy, further eroding the credibility of their security claims. This ruling by Judge Rita Lin reinforces First Amendment protections for companies in government contracting, limiting the executive branch's ability to use national security designations for retaliatory purposes. It signals that courts will meticulously scrutinize government justifications for such actions, particularly when protected speech is involved, thereby safeguarding the integrity of First Amendment AI government contracts.

This ruling reinforces First Amendment protections for companies in government contracting, limiting the executive branch's ability to use national security designations for retaliatory purposes. Lawyers should advise clients that courts will scrutinize government justifications for such actions, particularly when protected speech is involved.

Source: Original reporting via Court News Service