AI leaders urge UN to adopt global risk standards, US official opposes new governance
OpenAI, Anthropic and Hugging Face executives called for international AI safety standards at a UN meeting, while a senior US adviser rejected the idea of new global governance.
The chief executives of OpenAI, Anthropic and Hugging Face told United Nations delegates that the rapid pace of artificial-intelligence development creates societal risks that require coordinated international action. Sam Altman urged the creation of "national and international" standards for measuring AI capabilities, assessing risks, implementing safeguards and ensuring human oversight. He also advocated for rapid incident-reporting protocols so that failures can be shared globally.
Anthropic co-founder Dario Amodei, who recently published an essay calling for a slowdown in AI development, said that poorly managed AI could threaten humanity and that stricter safety standards are needed. He emphasized that managing these risks is an industry-wide, global responsibility.
Hugging Face CEO Clement Delangue echoed the call for common evaluation criteria, adding that coordinated standards would help countries compare evidence and verify compliance.
In contrast, Michael Kratsios, a senior technology adviser to the United States, told the forum that while AI development is accelerating and poses risks, those concerns do not justify pausing progress or establishing new global governance structures. He warned that international dialogue should not drift toward global regulation.
The differing positions highlight a split between leading AI firms seeking coordinated safety frameworks and a US official emphasizing competitive advantage and limited regulatory intervention.