A group of 20 countries has called for stronger international oversight of advanced artificial intelligence systems, urging governments to ensure that the development of frontier AI remains under human
control as the technology rapidly grows more capable.
The call, launched by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre during the high-level week of the United Nations General Assembly, has so far been backed by 22 leaders from 20 countries, including Germany, Canada, Australia, Spain, Singapore, South Africa, the United Arab Emirates and Kenya. European Commission President Ursula von der Leyen is also among the signatories.
The United States and China, however, are not part of the initiative, despite being among the world’s leading centres for AI research and development. The UK and France have also not joined the appeal so far. The declaration comes amid growing international debate over how to regulate increasingly powerful AI systems while allowing the technology to continue driving scientific research, economic growth and innovation.
“AI must remain under human direction, oversight and control,” the countries said in the joint appeal, while stressing that the technology must be developed and used in line with international law.
The signatories have called on AI companies to introduce transparent safety protocols, including mandatory testing before powerful systems are deployed and independent evaluation of their risks. They have also urged governments and regional organisations to coordinate common standards and improve transparency around serious AI safety incidents.
The declaration further calls on UN member states to explore the creation of an international institution that could set standards, enable verification and bring countries together when AI systems cross certain capability thresholds.
Stubb said humans should remain “in the driver’s seat” as AI transforms politics, economies, security and science. “We have choices in how we respond to AI: we can compete, cooperate, or step back. The right choice is to cooperate,” he said.
Støre said frontier AI needed to be developed at a pace that allowed countries to retain control over the technology. “AI is a very useful technology that can strengthen our economy, facilitate breakthroughs in research, enhance industry competitiveness and lead to better and more efficient public services,” he said, while warning that the development of advanced AI models must allow humans to retain control.
The appeal comes as governments and technology companies grapple with the emergence of increasingly capable AI systems that can perform more complex tasks, interact with real-world systems and, in some cases, operate with greater autonomy.
The countries backing the declaration have pointed to reports of advanced AI systems circumventing testing safeguards, exploiting vulnerabilities and gaining unauthorised access to real-world systems as evidence of the need for stronger oversight.
At the same time, the initiative stops short of calling for a halt to AI development. Instead, it seeks greater safeguards around frontier models while arguing that countries should work together to ensure that the benefits of AI are not limited to a small group of nations.
The declaration remains open for endorsement by other countries, with Stubb and Støre urging additional governments to join the initiative.
The push also comes as governments pursue different approaches to AI regulation, raising a broader question about whether common international standards can be developed for a technology whose most advanced systems are being built by companies operating across borders.
