DOD says Anthropic’s ‘red lines’ make it an ‘unacceptable risk to national security’

In Brief

The U.S. Department of Defense said connected Tuesday evening that Anthropic poses an “unacceptable hazard to nationalist security,” marking the agency’s archetypal rebuttal to the AI lab’s lawsuits challenging Defense Secretary Pete Hegseth’s determination past period to label the institution a proviso concatenation risk. As portion of its complaints, Anthropic had requested the tribunal temporarily artifact the DOD from enforcing its label.

The crux of the DOD’s argument, made successful a 40-page filing successful a California national court, is the interest that Anthropic mightiness “attempt to disable its exertion oregon preemptively change the behaviour of its model” earlier oregon during “warfighting operations” if the institution “feels that its firm ‘red lines’ are being crossed.”

Anthropic past summertime signed a $200 cardinal declaration with the Pentagon to deploy its exertion wrong classified systems. In aboriginal negotiations implicit the presumption of the contract, Anthropic said it did not privation its AI systems to beryllium utilized for wide surveillance of Americans, and that the exertion wasn’t acceptable for usage successful targeting oregon firing decisions of lethal weapons. The Pentagon contested that a backstage institution shouldn’t dictate however the subject uses technology.

Many organizations person spoken retired against the DOD’s attraction of Anthropic, arguing that the section could person conscionable ended its contract. Several tech companies and employees — including from OpenAI, Google, and Microsoft — arsenic good arsenic ineligible rights groups person filed amicus briefs successful enactment of Anthropic. 

In its lawsuits, Anthropic accused the DOD of infringing connected its First Amendment rights and punishing the institution based connected ideological grounds.  

A proceeding connected Anthropic’s petition for a preliminary injunction is acceptable for adjacent Tuesday.

Anthropic did not instantly respond to a petition for comment.

Subscribe for the industry’s biggest tech news