AI Governance Split: US Rejects Global Regulator While AI Leaders Call for Coordination
Updated 5:35 AM EDT, September 24, 2026
Michael Kratsios, director of the White House Office of Science and Technology Policy, told the UN Security Council in New York on September 23 that the United States opposes global governance of advanced AI,according to aTNWreport.
He referred to advanced AI as “superintelligence” throughout his remarks and said its rapid progress is not a reason to pause it or constrain it with new global governance structures.
He pointed to the Carolina Principles, which call on countries to preserve national control over how they regulate emerging technologies, while also backing voluntary international cooperation, best-practice sharing, and international standards developed through standards bodies.
The Carolina Principles were adopted as consensus principles by G20 innovation ministers in Chapel Hill, North Carolina, on September 2.
What Kratsios Rejects Global Governance
According toTNW, Kratsios said Washington remains open to dialogue on emerging technology, while rejecting global governance or control of superintelligence.
He cited President Trump’s remarks to the General Assembly the day before, saying the US “totally rejects any attempt to construct a global scheme of control of superintelligence.”
Kratsios also pointed to an AI exports program as one part of what the US offers other countries: help for partners in accessing secure AI infrastructure, systems, and services. “If you take the risks and benefits of superintelligence seriously, the most important thing you can do is to develop your country’s ability to understand and use it,” he said.
He argued the Security Council and similar bodies should focus on sharing best practices to build that capacity, and said, “A prosperous future will not be secured by a global regulator.”
The four briefers at the same meeting each proposed different specific measures, according toTNW.
Hugging Face’s CEO and co-founder Clément Delangue called for global monitoring and incident-disclosure standards, including mandatory sharing of AI agent traces. Yoshua Bengio, a Canadian computer scientist, and a pioneer of artificial neural networks and deep learning, urged a licensing approach for frontier AI in critical sectors, alongside liability insurance and incident reporting. Anthropic’s Dario Amodei pushed for international testing, verification, and notification standards, along with narrowly targeted restrictions on specific dangerous uses. OpenAI’s Sam Altman raised the idea of common benchmarks for assessing AI capabilities and safeguards.
Kratsios’s remarks did not reject international engagement broadly. The Carolina Principles he cited still support voluntary cooperation and shared standards work, even as he opposed global governance arrangements more broadly.
What This Means for Multinational Data Strategy
For chief data officers (CDOs), these differing positions make the direction of globalAI governanceuncertain.
It seems unlikely that international AI rules will converge into one harmonized framework soon. Building AI governance around a single anticipated global standard carries real execution risk while national and voluntary-international approaches continue to develop in parallel.
That said, the four briefers put forward several distinct proposals, from licensing to incident disclosure to shared benchmarks, at the same meeting where the US pushed back on global AI governance more broadly.
These are different governance mechanisms, and it isn’t yet clear which of them, if any, will gain real traction.
Related Stories
Similar Topics
Artificial IntelligenceData ManagementDiversityTestimonials
Shape the Data Leadership Agenda
Exchange Knowledge & Experience
About
Community
Events
Partner with Us
Contact Us
Submit Content
Join Our CommunityPrivacy
