Faced with demands from the US Department of Defense to allow its technology to be used for purposes the company considers unsafe or antidemocratic, Anthropic CEO Dario Amodei’s stance remains firm: “We cannot in good conscience accede to their request,” he wrote in a statement published on the company’s website late Thursday.
Anthropic is a significant US defense contractor: Its models are used in intelligence analysis, operational planning, and cyber operations, and it was one of the first to supply frontier AI models for use on US government “classified” networks.
However, Anthropic insists that its AI technology not be used to build autonomous weapons nor to conduct mass surveillance of the US population.
The DoD does not accept these restrictions and has said that, in future, it will only work with AI companies allowing “any lawful use” of their technology. If Anthropic does not agree to these terms then the DoD has threatened to designate the company a “supply chain risk” (which would bar it from participating in government contracts) and to force the company to work on the government’s terms by invoking the Defense Production Act.
“These latter two threats are inherently contradictory: One labels us a security risk; the other labels Claude as essential to national security,” Amodei wrote on Thursday.
Earlier in the week, Amodei met with US Secretary of Defense Pete Hegseth, who prefers the title Secretary of War.
Hegseth reportedly gave Anthropic until 5pm Eastern Time on Friday, February 27, to agree to his terms.