OpenAI Inks Deal with Pentagon
OpenAI has secured a landmark deal to deploy its AI models on classified U.S. Department of Defense networks. To address ethical concerns, the company proactively shared contract language explicitly forbidding its tech from being used for autonomous weapons or mass domestic surveillance.
This deal is part of a broader push by the Pentagon to accelerate AI adoption, managed by the Chief Digital and Artificial Intelligence Office (CDAO). The CDAO's new chief is Cameron Stanley, who previously led Project Maven, the military's flagship initiative for integrating AI into targeting and surveillance, from 2021 to 2022. The $200 million contract is for a pilot program to prototype "frontier AI capabilities" for both administrative and warfighting applications. Specific non-combat uses mentioned include improving healthcare for service members and streamlining the review of acquisition data. The work is centered in the Washington, D.C. region with a projected completion date of July 2026. This partnership follows a key change in OpenAI's usage policy in early 2024, when it removed a blanket prohibition on "military and warfare" applications. The company's Vice President of Global Affairs, Anna Makanju, who previously served as a senior policy advisor in the Obama/Biden administration, has been central to shaping OpenAI's policy and government engagement. OpenAI states the agreement includes technical safeguards to enforce its "red lines." These include a cloud-only deployment that prevents the models from running on edge devices, retaining full control over a "safety stack" to refuse certain tasks, and having cleared OpenAI personnel in the loop for oversight. The Pentagon is not working exclusively with OpenAI. The CDAO has awarded similar contracts of up to $200 million each to competitors including Google, Anthropic, and Elon Musk's xAI, fostering a competitive environment among top AI labs. This strategy aims to integrate the best commercial solutions to maintain a strategic advantage. This direct collaboration with AI developers like OpenAI marks a strategic shift for the Pentagon, which historically worked through traditional defense integrators. It mirrors the model pioneered by companies like Palantir, which took over as a key contractor for Project Maven after Google withdrew from the program in 2018 following employee protests. Other tech firms are also deepening their military ties. Palantir and Anduril Industries formed a consortium in late 2024 to merge their respective AI software platforms, the Maven Smart System and Lattice, to enhance data integration and targeting capabilities for the DoD.
