Pentagon threatens to blacklist Anthropic over AI restrictions

The U.S. Department of Defense has issued an ultimatum to artificial intelligence developer Anthropic, demanding the company allow its AI models to be used for all lawful purposes or face the termination of their partnership. Sean Parnell, the Pentagon’s chief spokesperson, stated that Anthropic has until 5:01 p.m. ET on Friday to decide on the request. “Otherwise, we will terminate our partnership with Anthropic and deem them a supply chain risk for DOW,” Parnell said. A supply chain risk is an official designation that can blacklist companies from government business. The Pentagon’s demand for lawful use—its justification for using AI models without vendor restrictions—seeks to bypass Anthropic’s internal safety policies. Anthropic refuses to allow its AI models to be used for mass surveillance of Americans or for fully autonomous weapons that conduct strikes without human input.

Defense Secretary Pete Hegseth has criticized Anthropic’s safety-first approach, characterizing it as woke AI and arguing that the military is building war-ready systems rather than chatbots. Hegseth asserted that the Department of Defense should not be limited by the rules of a vendor and that any lawful use of the technology should be permitted. In response to concerns, Parnell stated on Thursday that the Pentagon has no interest in conducting mass domestic surveillance or deploying autonomous weapons. However, Anthropic CEO Dario Amodei publicly signaled on Thursday that the company is not backing down, despite threats that it could be designated as a supply chain risk. The company’s policies specifically block its models from being used for the mass surveillance of American citizens or for fully autonomous weapons that execute strikes without human input.

If the Pentagon follows through on its threat, the consequences for Anthropic could be severe. Sachin Seth, a venture capitalist at Trousdale Ventures, reported that a supply chain risk label could mean the end of the company. However, dropping Anthropic could also create a national security gap for the U.S. military. Seth reported that the Department of Defense would have to wait six to 12 months for alternative AI providers like OpenAI or xAI to catch up if Anthropic is dropped. During this window, the military might have to rely on less capable models while waiting for these alternatives to reach the necessary capability levels.

Why it matters

The dispute over AI usage policies between Anthropic and the Pentagon threatens to blacklist the company from government contracts. This standoff highlights the growing tension between commercial AI safety standards and military operational demands, potentially forcing the U.S. military to rely on less capable models while waiting for alternatives to catch up.