OpenAI shares language from its contract with the Department of War

March 1, 2026 - 07:05

In a move to address growing public and industry concerns, OpenAI has publicly shared specific language from its contract with the U.S. Department of Defense. The disclosed clauses explicitly prohibit the use of its artificial intelligence technology for two highly sensitive applications: mass domestic surveillance and the development of autonomous weapons systems.

The company stated that these contractual safeguards are a core part of its commitment to the safe and ethical deployment of its AI models. By embedding these restrictions directly into its agreement, OpenAI aims to create a legally binding framework that governs how its tools, such as ChatGPT, can be utilized by defense agencies. This transparency effort appears designed to balance collaboration with government entities on projects deemed beneficial—like cybersecurity and veteran healthcare—with clear public assurances.

The disclosure comes amid intense global debate over the role of powerful AI in military and surveillance contexts. While the company has previously acknowledged working with the Defense Department on specific initiatives, releasing the precise contractual language marks a significant step toward operational transparency. OpenAI emphasized that all its government work remains subject to its usage policies, which forbid causing harm or developing weapons, but this contract solidifies those principles for its defense partnership specifically.