Building standards for the next phase of AI

OpenAI is calling for the United States to lead the development of global technical standards for frontier AI development, particularly for automated AI research systems that could improve themselves recursively. The organization argues that international standards are needed to prevent fragmentation across nations, enable collective action on shared risks, and address uneven capacity for AI development worldwide. Such standards would provide common definitions for safety practices, capability measurement, and risk assessment, while allowing individual nations to decide how to incorporate them into their own laws. OpenAI emphasizes that fully autonomous self-improvement systems are not happening today and should not be pursued until they can be done safely while maintaining human control and oversight.