OpenAI Urges U.S. to Lead International Standards for Frontier AI

OpenAI on Monday published a blog post urging the United States to take the lead in developing international technical standards for frontier AI, responding to growing concerns about the potential harms of artificial intelligence.

The company said international standards are necessary to safely advance AI, including defining which types of incidents involving AI models should be reported, clarifying how countries can cooperate on AI development, and improving access to the computing power needed to build such technologies. “Making continued AI progress safe and beneficial will require both advancing alignment research and developing shared standards that guide AI development across labs and countries,” OpenAI said.

OpenAI also said its CEO Sam Altman plans to speak at the United Nations Security Council on Wednesday on AI and international security. Over the weekend, U.S. Treasury Secretary Bessent said the United States has proposed a notification mechanism covering AI incidents with national security implications.

OpenAI emphasized that countries should develop complementary national and international standards for the most capable AI models, and suggested advancing the work through existing AI safety institutions in multiple countries, including the U.S. AI Safety and Innovation Center (CAISI) within the Department of Commerce. The company also called for standards addressing AI’s growing autonomous capabilities, saying recursive self-improvement with “full autonomy” — where AI could modify its own code or enhance its capabilities without human intervention — should not be pursued “until and unless safety is assured.”