OpenAI proposes U.S.-led global standards for frontier AI research

OpenAI proposed that the United States lead an international effort to create technical standards for frontier AI, including systems that automate work on successive generations of AI. The company wants common methods for measuring capabilities, supervising automated research and classifying and reporting incidents. Its suggested mechanism would build on the Center for AI Standards and Innovation and the international network of AI safety institutes, with participation from developers, researchers and standards bodies. OpenAI said the standards would not themselves license models or require pre-release approval, leaving governments to decide whether to incorporate them into law.