The US Calls for G20 to Avoid Creating AI Governance Institutions
September 1, 2026 – 2:37 pm
The United States will request the G20 not to establish any new institutions dedicated to governing artificial intelligence (AI), a White House official revealed ahead of a two-day meeting that commenced in North Carolina on Tuesday.
According to Reuters, this argument is procedural rather than substantive, meaning Washington isn’t proposing specific rules; they are urging member countries not to create bodies that could independently develop regulations.
The session, the first in a series leading up to the G20 leaders’ summit in Miami in December, sets the stage for negotiations over the coming months. The guest list includes prominent tech figures like Sam Altman and Jensen Huang, with Elon Musk participating virtually.
Canada has expressed support for regulation, advocating for a balance between AI innovation and public trust and safety. China’s attendance at the meeting remains unclear.
While the G20 lacks enforcement power, its agreed language often influences national regulatory actions. The US previously made similar arguments at the Paris AI Summit, where JD Vance cautioned against overregulating AI, and the country declined to sign an international statement on regulated AI use.
The current forum and stakes differ from previous discussions. A G20 process yields communiqué language that guides national regulators’ actions, making agreed-upon international oversight harder to modify than a summit speech.
Europe, already equipped with regulations like the AI Act, Digital Services Act (affecting platforms like ChatGPT), and Cyber Resilience Act, arrives at this meeting from a different perspective. The timing is significant as the Financial Stability Board’s chair warned G20 finance ministers that AI-driven cyber risk poses an immediate threat to financial stability, highlighting through such multilateral bodies the US wants left untouched.
A UN panel has also cautioned that AI development outpaces both scientific understanding and government policy. Yet, European regulators face a practical dilemma: while nothing agreed in Miami would undo the AI Act, a G20 consensus against international oversight could make Europe’s enforcement efforts appear distinct rather than exemplary.
The counterargument to the US’ stance deserves careful consideration. Establishing institutions now would be challenging given the current lack of understanding about AI, staffing issues, and the need for constant adaptation to evolving technologies. There’s also a competitive element: countries hosting prominent AI model developers stand to gain less from international oversight than those without significant AI presence. Developing nations have distinct interests and concerns in this context.