OpenAI has called for the United States to lead an international effort to develop technical standards for frontier artificial intelligence, Reuters reported on September 21 as world leaders gathered for the United Nations General Assembly.
In a blog post, the company said governments and technical bodies should develop complementary national and international standards for advanced AI systems, including common measurements and incident-reporting protocols. Reuters said OpenAI specifically included the possibility of recursive self-improvement, in which an AI system can autonomously improve aspects of its own capabilities.
The proposal is a policy position from OpenAI, not an adopted U.S. government or international standard. The company argued that U.S. leadership could help shape a more consistent global framework as increasingly capable systems are developed in multiple countries.
AI safety and governance are also on the agenda at the United Nations this week. OpenAI chief executive Sam Altman is due to brief the U.N. Security Council, while industry leaders and governments continue to disagree over how quickly advanced AI should be developed and how risks should be managed.
The debate remains unsettled, and no binding global technical regime has been agreed.