Anthropic, a prominent artificial intelligence American firm, is challenging a government order that deemed it a national security threat for not granting unfettered access to its Claude AI model to the military. The company filed an appeal Tuesday in a San Francisco federal court requesting an injunction against the designation, however, because the restriction could be punitive.
At the beginning of the hearing, District Judge Rita F. Lin was concerned that the government’s action “seems to be a punishment for Anthropic’s public disagreement with the U.S. government.” In February, Trump and Defense Secretary Pete Hegseth announced that Anthropic would no longer assist the Pentagon in the development of lethal autonomous arms or mass surveillance of Americans after it rejected the proposal.
The US government declared Anthropic a “supply chain risk to national security” and ordered federal agencies to discontinue their use of the AI model. Anthropic in turn sued on March 9 – both over the supply chain risk designation and for the Trump administration infringing upon the company’s First Amendment rights.
Judge Lin agreed that the Pentagon can decide which AI tools to use, but wondered if banning all federal use and publicly urging contractors to sever ties with Anthropic was a step too far. Government attorneys said the move was only in response to the “potential security risk” that Claude AI represents, not as a response to Anthropic’s public comments. They also pointed out that future improvements to Claude may raise further security issues.
The case has attracted attention from industry experts, who see the ruling as potentially precedent-setting. It’s unusual for a domestic artificial intelligence company to be given a supply chain risk label and would give the executive branch wide discretion to redefine laws, Ben Goertzel, computer scientist and CEO of SingularityNET, told Euronews Next. If the designation prevents Anthropic from selling software to other companies that work with the government, the consequences could be serious, “but there will be ways for Anthropic to work in commercial markets,” he said.
Goertzel also said that the ban could have a broader impact on the AI sector, putting pressure on other firms to bow to government demands. “He’s [President Trump] trying to somehow force the AI industry into line — like everyone else,” Goertzel said.
Judge Lin said that she will decide in the next few days whether to temporarily halt the government ban pending a comprehensive review of the case. The ruling could help dictate the future of government and military involvement in American AI companies and limit the extent of governmental control over AI in commercial activities.
Anthropic was unable to comment at the time of writing. The case remains under close scrutiny by AI developers and legal scholars who are interested in the intersection of technology, national security, and free speech.