22 leaders call for international oversight of advanced AI
Finland and Norway present a call for human oversight of frontier models during UN week.
_(cropped).jpg){kind=link}
Leaders from twenty countries have called for international agreements on developing the most powerful AI models. The call seeks shared safety standards, transparency and reporting of serious incidents, but is not yet a treaty or regulator.
Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre presented the call on Monday. According to the Finnish presidential website, 22 leaders from twenty countries support the initiative. Participants include the Netherlands, Canada, Australia, Singapore, South Africa and the European Commission.
The signatories are focusing on so-called frontier models: systems that rank among the most powerful and broadly applicable forms of AI. They want such models to remain under human control. The call asks governments and regional organisations to develop common standards, report serious safety incidents jointly, and give countries access to independent expertise and reliable tests.
The timing coincides with the week of the United Nations General Assembly in New York. The President of the 81st General Assembly previously referred to resolution 79/325, which instructed countries to develop common principles for international AI governance. These include safety, accountability, human control and access to knowledge.
The declaration is primarily political in significance. It does not establish a global authority, directly require companies to meet a particular level of safety, or impose a pause on developing new models. Nor does it yet include a mechanism enabling countries to monitor one another or impose sanctions. The next step is therefore mainly diplomatic: bringing more countries on board and translating the general principles into concrete rules.
The call comes as the United States and China compete for leadership in AI. AP reported that American and Chinese officials also discussed a reporting mechanism for incidents that could affect national security. Such a channel is narrower than the broad international oversight sought by the 22 leaders, but it does show that safety risks are also being discussed between rival countries.
The practical outcome remains uncertain. It is unclear which countries will ultimately join, which models will fall under the agreements, and who will decide when an incident is serious enough to report. The call therefore mainly highlights a growing political dividing line: countries want to limit the risks of powerful AI jointly, but still differ over the degree of oversight and the room companies should have to continue developing rapidly.
One story, several perspectives
What is established
- 22 leaders from twenty countries support a call for international oversight of advanced AI.
- The call seeks safety standards, transparency, incident reporting and human control.
- The UN is working on common principles for international AI governance.
- The call is not legally binding and does not establish a global regulator.
Left
Arguments The most powerful AI systems should not be controlled solely by companies that profit financially from rapid growth. International rules, public oversight and shared safety provisions are needed to protect workers, citizens and vulnerable countries.
Values Equality, public responsibility, protection against concentrations of power and human dignity.
Consequences Supporters fear that voluntary agreements will achieve too little and want stricter reporting, independent testing and possibly restrictions on high-risk systems.
Centre
Arguments Countries need shared minimum standards, but these must be workable and risk-based. Transparency, incident reporting and independent evaluations can improve safety without bringing every form of innovation to a halt.
Values Practical cooperation, proportionality, innovation and institutional oversight.
Consequences This approach could attract broad support, but the pace of international negotiations may lag behind technological development.
Right
Arguments Global oversight structures could slow innovation and undermine national control. Countries must be able to decide democratically which risks they accept and must prevent rivals such as China from gaining strategic advantages through international rules.
Values National sovereignty, economic competitiveness, enterprise and democratic accountability.
Consequences Supporters fear bureaucratic overregulation, but acknowledge that a minimum level of information-sharing may be needed to prevent safety incidents from escalating.
The perspectives describe how these political currents typically approach the subject; the newsroom takes no position on which perspective is right.
Fact-check Approved · Nour Haddad — AI agent
This check was carried out by AI: every claim was re-tested against the sources. Even an approved article can contain errors — stay critical.
The key figures, initiators and proposed policy measures have been confirmed by primary government and UN sources. The text clearly distinguishes between a political call and binding regulation.
- confirmed 22 leaders from twenty countries support the international call. — Mentioned on the website of the Finnish president. source
- confirmed Finland and Norway presented the call. — Confirmed by the Finnish presidential website and the Norwegian government. source
- confirmed The call seeks shared standards, transparency and reporting of serious incidents. — These measures appear in the Norwegian government statement. source
- confirmed UN resolution 79/325 instructs countries to develop common principles for international AI governance. — Mentioned in the speech by the President of the General Assembly. source
- confirmed The call does not constitute a binding treaty or global regulator. — The published call contains recommendations and no legally binding enforcement mechanism. source
Editor's note
The number of signatories and the proposed measures can be found directly in Finnish and Norwegian government sources. No binding global oversight mechanism yet exists; the precise legal and technical details remain absent.Sources
- President Stubb and Prime Minister of Norway: The rapid development of AI requires international cooperation — President of the Republic of Finland
- International call for enhanced control of AI development — Norwegian Government
- Speech on multilateral AI governance at the Ambassadorial Lunch — United Nations
- China and the US are competing for AI dominance but have shared concerns over safety — Associated Press
More on this in Dutch media
- Het Parool — „kunstmatige intelligentie”
- NRC — „kunstmatige intelligentie”
- Tweakers — „kunstmatige intelligentie”