OpenAI has reached an agreement with the U.S. Department of War to deploy its artificial intelligence models on classified networks, marking a major milestone in AI integration with national defense systems.

The deal comes amid rising tensions over AI safety and Pentagon contracts.

OpenAI CEO Sam Altman announced the deal on X (formerly Twitter), stating that the Department of War displayed “a deep respect for safety” and a shared desire to achieve “the best possible outcome.”

Altman confirmed that OpenAI’s models will operate under strict restrictions, including prohibitions on domestic mass surveillance and requirements for human oversight in the use of force, including autonomous weapon systems.

“Two of our most important safety principles are reflected in law and policy, and we incorporated them into our agreement,” Altman said.

Anthropic excluded from Pentagon deployment

The agreement follows a week of tension in the AI sector. Defense Secretary Pete Hegseth labeled Anthropic a “Supply-Chain Risk to National Security,” effectively halting its deployment on classified networks.

Anthropic had sought assurances that its models would not be used for fully autonomous weapons or domestic surveillance, but the DoW rejected its conditions. Anthropic expressed deep disappointment and plans to challenge the designation in court.

OpenAI promises technical safeguards

Altman assured that OpenAI will implement technical safeguards and deploy personnel to oversee model safety. He also urged the Department of War to extend similar terms to other AI companies to foster “reasonable agreements” over legal or governmental conflicts.

The move distinguishes OpenAI from competitors and reinforces its reputation for balancing AI innovation with safety standards acceptable to the U.S. military.