The D.C. Circuit Court of Appeals has upheld the U.S. Department of Defense's ban on the use of Anthropic's Claude models. The court ruled 2-1 that the Pentagon had sufficient grounds to designate the technology as a national security risk under the 2018 Federal Acquisition Supply Chain Security Act (FASCSA).
The conflict arose after Anthropic refused to modify its user agreement to allow all lawful uses of the AI. The company's existing terms prohibit mass surveillance and the development of autonomous weapons. Defense Secretary Pete Hegseth implemented the ban in June, restricting access for both Pentagon staff and defense contractors.
Anthropic previously secured a $200 million contract in July to provide Claude models to the Pentagon, with plans to integrate them into the GenAI.mil platform. The company argued that the Department of Defense failed to follow proper notification procedures required by FASCSA, but the court rejected this claim.
The legal landscape remains complex, as a separate ruling by the U.S. District Court for the Northern District of California in August declared the ban illegal under the 2011 National Defense Authorization Act. Anthropic stated it disagrees with the latest decision and is evaluating options, including potential appeals to the Supreme Court.
