Anthropic has turned down a fresh proposal from the U.S. Pentagon, which was intended to break a deadlock between the two parties. Tensions had arisen previously due to a disagreement over the terms under which Anthropic’s artificial intelligence software could be used for military purposes. This confrontational approach had put Anthropic’s government-related defense projects at risk. A spokesperson for Anthropic clarified that the Pentagon’s newly proposed terms fell short of the company’s standards. The Department of Defense had earlier set a Friday deadline for Anthropic to accept the government’s conditions, warning that failure to comply would result in the company being flagged as a supply chain risk. In response, Anthropic’s CEO emphasized that such threats would not sway the company’s decision. He further mentioned that Anthropic had suggested a collaborative research and development effort with the Department of Defense to improve system reliability—a proposal that was ultimately declined by the Pentagon.