US judge says Pentagon’s Anthropic blacklisting looks like punishment for AI safety views
U.S. District Judge Rita Lin in San Francisco, an appointee of former Democratic President Joe Biden, said during the hearing that the appointment “appears to be an attempt to cripple Anthropic.”
Also Read | OpenAI sweetens private equity pitch amid corporate turf war with Anthropic
“DOW appears to be punishing Anthropic for trying to bring this contract dispute to public scrutiny,” Lin said, using the acronym for War Department, President Donald Trump’s new name for the Department of Defense.
The hearing concerns Anthropic’s request to temporarily block the appointment while litigation continues. The unprecedented appointment, which comes after Anthropic refused to allow the military to use its Claude AI software for US surveillance or autonomous weapons, prevents Anthropic from exiting certain military contracts. This could cost the company billions of dollars this year in lost business and reputational damage, Anthropic executives said March 9.
The company opposes seeing domestic surveillance as a violation of rights, stating that artificial intelligence models are not reliable enough to be used safely in autonomous weapons.
THE FIRST ANTHROPIC DESIGN FOR OUR USA COMPANY
Anthropic’s designation was the first time a U.S. company was publicly identified as a supply chain risk under an obscure government procurement law aimed at protecting military systems from foreign sabotage. In its lawsuit filed March 9, Anthropic alleged that the government violated its right to free speech under the First Amendment to the Constitution by retaliating for its views on AI safety. The company said it was not given a chance to challenge the appointment, violating its Fifth Amendment right to due process.
Also Read | Antropik challenges Pentagon’s national security risk claim in response to lawsuit
The lawsuit says the decision was unlawful, unsupported by facts, and inconsistent with the military’s past praise for Claude.
The Justice Department countered that Anthropic’s refusal to lift the restrictions could lead to uncertainty at the Pentagon about how it could use Claude and pose a risk of disabling military systems during operations, according to the court filing.
The government said the appointment was due to Anthropic’s refusal to accept contract terms, not its views on AI security. Anthropic has a second lawsuit pending in Washington, D.C., over a separate supply chain risk designation that could lead to the Pentagon removing it from civilian government contracts.
