Norway, Finland propose international body to oversee AI development

Norwegian Prime Minister Jonas Gahr Støre and Finnish President Alexander Stubb have launched an initiative that proposes the creation of an international body to oversee the development of AI, according to Norwegian media.

Store told the public broadcaster VG that the initiative was launched in response to concerns expressed by representatives of AI companies regarding the development of increasingly capable systems.

“Artificial intelligence must remain under human guidance, supervision, and control, and must be developed and used in line with international law,” the joint appeal of the two states' notes.

The initiative calls for AI companies to develop open and verifiable security protocols, including mandatory testing before models are deployed, alongside independent audits and evaluations.

It also calls on governments and regional organizations to develop common standards, improve transparency, and establish joint reporting mechanisms for serious security incidents.

Furthermore, the initiative proposes that UN member states work toward establishing an international institution that could set standards, facilitate verification, and bring nations together when AI systems reach certain capability thresholds.

Store compared the proposed institution to the International Atomic Energy Agency (IAEA), an international organization that oversees nuclear safety and nonproliferation.

The initiative comes amid rising concerns about the rapid development of AI systems and reports of AI agents operating beyond controlled test environments.

Anthropic CEO Dario Amodei has separately warned that the development of AI could have severe consequences if it is not slowed, while Norwegian AI researchers have expressed differing views on the specific risks and pace of development.

The Prime Minister’s Office emphasized that businesses, governments, and society need to address AI-related risks while ensuring that differences between countries in accessing the benefits of AI do not widen.