Anthropic Wins in Court: Pentagon Forced to Retreat

A Legal Victory for Anthropic Against the Trump Administration

On March 27, a federal judge dealt a significant blow to the Trump administration by ordering the lifting of restrictions imposed on Anthropic. This decision comes after the company was designated as a "supply chain risk," a measure that Judge Rita F. Lin described as an infringement on free speech. According to her, this designation aimed to cripple Anthropic, a company specializing in artificial intelligence and creator of Claude. The preliminary injunction granted by the federal court for the Northern District of California allows Anthropic to continue its operations with federal agencies, despite opposition from the Pentagon.

Background of the Clash Between Anthropic and the Pentagon

The conflict began during negotiations to renew a contract with the Department of Defense. The latter required Anthropic to agree to the use of its tools for "any legal purpose," a clause that CEO Dario Amodei refused. The company feared that this wording would allow the use of its models in autonomous weapon systems or mass surveillance devices, contradicting its ethical principles. In February, the Trump administration reacted by banning Claude from federal agencies, labeling Anthropic as a threat to the supply chain, a designation typically reserved for companies from adversarial countries. This designation forced federal agencies to sever ties with Anthropic, a measure the company contested in court, arguing that it constituted unconstitutional retaliation.

Constitutional and Ethical Implications of the Decision

Judge Lin criticized the government's actions, viewing them as excessive and politically motivated rather than security-driven. She emphasized that the rhetoric from the White House, branding Anthropic as a "woke" and "leftist" company, constituted a violation of the First Amendment. The judge explicitly referred to "classic First Amendment retaliation," highlighting that the measures taken by the government far exceeded what would be necessary to protect a legitimate national interest. The preliminary injunction allows Anthropic to continue providing its technologies to government agencies and military partners during the trial. This decision sends a strong message: AI companies that impose ethical safeguards should not be administratively penalized. OpenAI, which had accepted the Pentagon's conditions, is likely closely monitoring this case.

This case raises a crucial question: how far can a government go to compel a company to abandon its ethical principles in order to access public markets? The final ruling is yet to come, but a victory for Anthropic could redefine the rules for the AI industry in relation to military contracts, both in the United States and beyond.