[OpenAI Releases 'National Security Principles,' Defining AI Usage Boundaries in Government and Defense] OpenAI has released its 'National Security Principles,' clarifying the boundaries and usage norms for its collaboration with governments in the fields of national security and law enforcement. These principles will apply to existing and future partnerships with organizations such as the U.S. Department of Defense, reaffirming the prohibition of using its technology for large-scale domestic surveillance, autonomous weapons system control, and high-risk automated decision-making. OpenAI stated that it has established 'trusted access' collaborations in the areas of cybersecurity and biosafety with the U.S. and its allies, including partnerships with multiple national cybersecurity agencies through the Daybreak program. Additionally, it has made its biosafety model, GPT‑Rosalind, available to certain government agencies for public health and biodefense missions. OpenAI expressed its support for legislation to impose external constraints on high-risk military AI applications.