A US federal court on Thursday granted a preliminary injunction blocking sweeping government restrictions against AI company Anthropic, finding the measures were likely unlawful and caused significant harm to the firm.

The US District Court for the Northern District of California ruled that Anthropic is likely to succeed in its claims that federal actions against it constituted retaliation for protected speech and violated due process rights.

The case centers on a dispute between Anthropic and the US Department of War over the use of the company’s AI model, Claude, in military and surveillance contexts.

Anthropic argued that its technology “is not ready for safe use in fully autonomous lethal weapons or the mass surveillance of Americans,” and sought to impose limits on such uses.

In response, federal officials announced a series of measures, including a government-wide ban on using Anthropic’s technology, restrictions on contractors working with the company, and its designation as a “supply chain risk.”

The court found these actions appeared punitive rather than necessary for national security, stating they “appear designed to punish Anthropic” and could “cripple” the company.

Judge Rita F. Lin ruled that the record supports an inference that Anthropic was targeted for “criticizing the government’s contracting position in the press,” calling the actions “classic illegal First Amendment retaliation.”

The court also found procedural violations, noting Anthropic had “no notice or opportunity to respond,” likely breaching due process protections.

Additionally, the ruling determined that labeling Anthropic a supply chain risk lacked a lawful basis and was “arbitrary and capricious,” with no clear evidence supporting claims it posed a threat to national security systems.

The decision highlighted that the government could have addressed concerns by simply discontinuing use of the company’s technology, rather than imposing broad restrictions affecting federal agencies and private contractors.

The court concluded that Anthropic is suffering “irreparable harm,” including loss of business opportunities and reputational damage, while the public interest favors protecting constitutional rights and open debate on AI safety.

The preliminary injunction temporarily blocks enforcement of the challenged measures, while allowing the government to seek an expedited appeal. (AA)