OpenAI Signs Landmark Pentagon Deal With Strict AI Guardrails, Banning Autonomous Weapons and Mass Surveillance

OpenAI strikes a landmark Pentagon deal deploying advanced AI in classified environments, banning autonomous weapons, mass domestic surveillance, and high-stakes automated decisions, while retaining full safety control and the right to terminate the contract if violations occur.

• OpenAI reaches a landmark agreement with the Pentagon for deploying advanced AI in classified environments, establishing stricter guardrails than any previous classified AI deployment, including explicit prohibitions on mass domestic surveillance, autonomous weapons direction, and high-stakes automated decisions.
• The deal operates exclusively through cloud-based deployment, with OpenAI retaining full control over its safety stack and cleared OpenAI personnel remaining actively involved, ensuring red lines cannot be crossed and allowing the company to terminate the contract if violations occur.
• OpenAI requests that the same contract terms be made available to all AI companies, calls for resolution of tensions between the Department of War and Anthropic, and confirms it does not support designating Anthropic as a supply chain risk, signaling a push for broader industry collaboration with the government.