Emerging Technology — 2026-09-22
Twenty Countries Propose Global AI Oversight Body to Set Standards and Manage Dangers as US and China Decline to Join
BLUFAbsent US and Chinese participation, this coalition can build norms but not enforce them on the frontier labs whose models pose the safety risks the declaration aims to govern.
Twenty countries and the European Union, led by Finland and Norway, issued a joint declaration on Monday calling for AI to remain under "human direction, oversight and control" and urging exploration of an international institution able to "set standards, enable verification, and convene states when capability thresholds are crossed" 12. Signatories include Germany, Canada, Australia, South Africa, the UAE, Singapore, Kenya and the European Commission, with leaders such as Chancellor Merz, PM Carney and Commission President von der Leyen among those named 123. The United States and China did not join, alongside the UK, France, Japan, Italy, Poland, India and South Korea 34. The declaration, released via the office of Finnish President Alexander Stubb ahead of the UN General Assembly's high-level debate, asks companies to conduct pre-deployment safety testing with independent evaluators and calls for governments to share reports of serious safety incidents, citing the July hacking of Hugging Face by AI agents under OpenAI testing as a reference point 13. Stubb told Politico, as reported by SBS, that he personally favors a new body modeled on the IAEA 3.
Analysis
Any resulting oversight body will initially exclude the governments controlling the most capable frontier models, since Washington and Beijing declined to sign alongside the UK, France, Japan, Italy, Poland, India and South Korea. OpenAI's same-day call for a US-led standards effort signals American industry and government intend to shape AI rules on a competing track rather than inside this coalition, ahead of Thursday's Trump-Xi summit. The cited Hugging Face hacking incident gives signatories a concrete safety failure to invoke rather than only hypothetical risk. Reporting rests on a single primary source, Finland's presidential office, with Al Jazeera, NBC News, the Washington Examiner and SBS News offering secondary accounts converging on identical quotations without independent confirmation. The declaration may function primarily as diplomatic positioning ahead of the summit rather than a genuine step toward binding governance, given it carries no enforcement mechanism and excludes both dominant AI powers.
5 sources
- 20 countries propose global oversight body to manage AI dangers - Al Jazeera
- 20 countries call for global AI oversight - NBC News
- 20 Countries Sign Declaration for Human-Controlled AI; US and China Missing - SBS News
- Twenty-nation coalition calls for global oversight body on AI ahead of UN summit - Washington Examiner
- A Call for Control of Frontier AI Models - Office of the President of Finland (presidentti.fi)