Our approach to government and national security partnerships

Governments are increasingly utilizing advanced AI systems for critical national security functions, necessitating collaboration among AI developers, governments, and civil society to define responsible deployment. OpenAI believes democratic societies can leverage AI to safeguard citizens, protect infrastructure, deliver public services, and counter threats in areas like cyber defense and biological security. However, AI implementation must uphold democratic accountability, human judgment, and the rule of law, reinforcing democratic institutions.

To promote transparency, OpenAI has published its National Security Principles, detailing its approach to government partnerships and national security applications of its technology. Developed through extensive internal and expert consultation, these principles guide the company's work with entities like the U.S. government and allied partners.

This initiative expands OpenAI's engagement with international partners, including those in the Daybreak cyber defense program and countries like Australia, Canada, Japan, and European Union institutions. The company is also applying a similar strategy to biosecurity, exemplified by the expanded access to its GPT-Rosalind model for U.S. government and allied partners in public health and biodefense.

These principles extend to all national security and law enforcement collaborations, including the company's work with the Department of War. Specific contractual limitations are in place, prohibiting the use of OpenAI's technology for mass domestic surveillance, autonomous weapons systems, or critical automated decisions. OpenAI advocates for transparency in AI usage and supports legislative efforts to establish safeguards for high-risk military AI applications, asserting that such decisions should be rooted in democratic processes.