OpenAI, Anthropic, and Hugging Face executives have called for international coordination to establish shared standards for AI capability evaluation, safety testing, and incident reporting at the United Nations. OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei emphasized that as AI capabilities grow, common frameworks are necessary to monitor risks and maintain human control over systems.
The companies propose creating unified protocols for measuring model performance and disclosing safety incidents, allowing different countries to compare evidence. OpenAI specifically advocates for the United States to lead the development of these international standards, ensuring that technical frameworks are interoperable rather than delegating regulatory authority to global bodies.
The Trump administration has formally rejected the creation of new global AI governance frameworks. Presidential science advisor Michael Kratsios stated that while the U.S. acknowledges the risks associated with AI, it does not believe these issues justify slowing down development or ceding regulatory power to international organizations.
This stance aligns with President Trump's September 22 statement at the UN General Assembly, which prioritized maintaining U.S. competitive advantage in AI. The administration supports technical coordination and dialogue but opposes any global governance structure that could impose binding rules or restrict the speed of American AI innovation.
The debate highlights a shift in AI safety from an internal corporate concern to a national policy issue. While industry leaders seek standardized risk metrics and cross-border incident reporting to manage potential threats, the U.S. government remains committed to preventing the formation of any international entity that could override domestic regulatory oversight.
