Anthropic filed two sworn declarations with a federal court in California, challenging the Pentagon’s claim that the AI company poses a national security risk. They argue the government’s case is based on technical errors and previously unmentioned claims.
The declarations accompanied Anthropic’s lawsuit reply brief against the Department of Defense, with a hearing set for Tuesday, March 24, before Judge Rita Lin in San Francisco.
The issue began in February when President Trump and Defense Secretary Pete Hegseth cut ties with Anthropic after it refused unrestricted military use of its technology.
Declarations were made by Sarah Heck, Anthropic’s Head of Policy, and Thiyagu Ramasamy, the company’s Head of Public Sector.
Heck, formerly with the National Security Council, disputed the claim that Anthropic demanded a role in military operations, stating it was untrue and pointing out the government’s new concerns were never raised during negotiations.
Heck highlighted a March 4 email from Under Secretary Emil Michael to CEO Dario Amodei, where Michael said they were “very close” on issues now cited as threats, questioning his later claims of no negotiations.
Ramasamy, with a background in AI deployments for government clients, refuted the idea that Anthropic could disable or alter technology in military use, explaining technical controls prevent such interference.
He also argued against claims that hiring foreign nationals is a risk, emphasizing Anthropic staff have undergone security clearances.
Anthropic’s lawsuit contends the security risk designation is retaliation against its AI safety views, violating the First Amendment. The government maintains it’s a security measure, unrelated to Anthropic’s opinions.
In another legal issue, a judge tentatively decided that Reddit’s lawsuit against Anthropic for unauthorized content scraping will return to state court, with a hearing also on Tuesday.