Strengthening Democratic Oversight in National Security

OpenAI is launching an initiative to help democratic oversight bodies develop expertise and tools to understand and supervise how governments use AI for national security purposes. This matters because AI systems can operate at speeds that make traditional manual oversight difficult, and mistakes can spread before being detected, yet oversight is especially important in national security where decisions involve secrecy and high-stakes consequences. The initiative is guided by three principles: AI should strengthen rather than replace human judgment, government AI use should be traceable to authorized reviewers, and AI tools should empower oversight institutions to scale their review capacity. Over the next year, OpenAI plans to work with officials to identify oversight opportunities, provide funding and credits to democratic oversight bodies, pilot tools to help reviewers examine AI-assisted decisions, and engage civil society in developing these approaches.