OpenAI and US Department of Defense Agree to Terms for AI Deployment in Military Settings

OpenAI has announced a deal with the United States Department of Defense (the “DoD”) that will allow the United States military to deploy advanced artificial intelligence systems within classified environments. The arrangement follows the recent breakdown in relations between Anthropic and the US government, which resulted in Anthropic being designated a “Supply-Chain Risk to National Security”, and federal agencies being directed to discontinue all use of Anthropic technology.

Under the new agreement, OpenAI and the DoD have established a series of guardrails designed to place clear limits on how military entities may use advanced AI. The safeguards include prohibitions on employing the technology for mass domestic surveillance, directing autonomous weapons systems, or making high-stakes automated decisions. The agreement also references DoD Directive 3000.09, which mandates that systems undergo “rigorous hardware and software verification and validation” to ensure that it performs as intended in realistic environments prior to deployment.

OpenAI CEO Sam Altman has told staffers that he is hoping to “de-escalate” tensions between Anthropic and the DoD, stating that both companies share the same redlines when it comes to the use of AI in military scenarios. A US$200 million contract is already in place between OpenAI and the DoD that has allowed the agency to use the startup’s models in non-classified use cases.

Author: Jack Borins, 2025-2026 Articling Student-At-Law