Strengthening democratic oversight in national security

OpenAI has announced a new initiative aimed at strengthening democratic oversight of artificial intelligence in national security applications. According to the company, while AI can assist democratic governments in stopping cyberattacks, securing critical infrastructure, and detecting threats, its speed and scale make traditional, paper-era manual oversight methods nearly impossible. To help oversight institutions keep pace, OpenAI is providing $5 million in training, technical support, and credits to support the responsible adoption and evaluation of AI tools. Over the next year, the initiative will involve working with authorized officials to identify practical oversight opportunities, piloting model-agnostic tools that help reviewers examine records surrounding AI-assisted government decisions, and engaging civil society and technical experts. OpenAI emphasizes that it will not perform an oversight role over the government, nor will it control findings; participating institutions will retain complete control of evidence and outputs. The company states that AI should strengthen human judgment, remain traceable to authorized reviewers, and empower democratic oversight bodies rather than replace them. Progress will be measured by whether authorized reviewers can better fulfill their obligations and whether the public gains stronger trust in government AI systems.