Pentagon says Anthropic’s ‘red line’ poses ‘unacceptable risk to national security’

The Pentagon said Tuesday night that Anthropic poses an “unacceptable risk to national security,” the department’s first response to the AI ​​Institute’s lawsuit challenging Defense Secretary Pete Hegseth’s decision last month to designate the company as a supply chain risk. As part of its complaint, Anthropic had asked the court to temporarily block the Department of Defense from enforcing the label.

At the heart of the Pentagon’s argument, laid out in a 40-page filing in federal court in California, is concern that Anthropic may “attempt to disable its technology or proactively alter the behavior of its models” before or during “combat operations” if it “feels that it has crossed a corporate ‘red line’.”

Anthropic signed a $200 million contract with the Department of Defense last summer to deploy its technology within classified systems. In subsequent negotiations over the terms of the deal, Anthropic said it did not want its AI system to be used for mass surveillance of American citizens and that the technology was not ready to be used to determine the targeting or firing of lethal weapons. The Pentagon argued that private companies should not decide how the military uses technology.

Many organizations object to the Pentagon’s treatment of Anthropic, arguing that the department could have terminated the contract. Several technology companies and employees, including OpenAI, Google, and Microsoft, as well as legal rights groups, filed court briefs in support of Anthropic.

In its lawsuit, Anthropic accused the Department of Defense of violating its First Amendment rights and punishing the company on ideological grounds.

A hearing on Anthropic’s request for a preliminary injunction is scheduled for next Tuesday.

Anthropic did not immediately respond to a request for comment.