Building standards for the next phase of AI

OpenAI has published an outline advocating for shared global AI standards to improve safety as artificial intelligence systems grow more capable and autonomous. According to OpenAI, its mission involves navigating future AI progress while ensuring systems remain aligned with human values and under human control. The organization points to the potential of automated AI research, which could drive recursive self-improvement and accelerate the pace of development. However, OpenAI cautions that fully autonomous recursive self-improvement should not be pursued unless it can be done safely, noting that a lack of robust safeguards could lead to humans losing practical control over AI development. To navigate these challenges safely, OpenAI calls for international standards for safety and security practices in frontier AI development. These standards would help address issues such as regulatory fragmentation across nations, collective action risks, and uneven global capacity in frontier development. OpenAI suggests that the United States should lead an effort to work with countries around the world to develop global technical standards for frontier AI, leveraging the emerging network of international AI safety institutes. The proposed standards would provide a common technical foundation for capability measurement, evaluation, risk assessment, and safeguard sufficiency, without acting as mandatory licensing or prerelease review requirements. Furthermore, OpenAI emphasizes that these standards should support a competitive ecosystem by consulting open and closed model developers, independent experts, and academia, drawing lessons from regulated sectors like aviation and financial stability.