20 Nations Call for Human Oversight of AI Systems

International Coalition Seeks AI Guardrails

Twenty countries have signed a declaration calling for stronger international oversight of artificial intelligence systems, with an emphasis on maintaining human control over AI development and deployment. The initiative, spearheaded by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre, notably lacks participation from the United States and China.

The declaration responds directly to recent security incidents, including a cyberattack involving rogue AI agents on the Hugging Face platform. It calls for AI companies to implement transparent safety protocols with mandatory pre-deployment testing and independent evaluation by qualified assessors who must be granted sufficient access to assess risks.

Why It Matters

As AI systems grow more powerful and autonomous, the absence of international coordination creates regulatory gaps that could enable dangerous deployments. This declaration represents an attempt to establish baseline safety standards before a major incident occurs, though its effectiveness depends on whether major AI powers eventually participate.

Proposed Governance Framework

The signatory nations are advocating for several concrete measures. Countries should develop and coordinate common standards while strengthening transparency requirements, including shared reporting of serious safety incidents. The declaration also urges UN member states to build on existing international mechanisms and explore creating a dedicated international institution for AI oversight.

Stubb indicated his personal preference would be establishing a body similar to the International Atomic Energy Agency within the UN framework. However, he acknowledged the need for flexibility in the institutional design, particularly given skepticism toward UN-based solutions from some quarters. "We have to leave it open-ended so that people don't reject it just because it would be within the U.N. framework," he told POLITICO.

Støre suggested that more focused forums like the G7 and G20 might provide effective venues for advancing the agenda, noting that the United Kingdom's upcoming G20 presidency could be particularly important.

What the Declaration Avoids

The document deliberately stops short of calling for any slowdown or pause in AI development. Organizers emphasized their focus on managing risks rather than constraining innovation. This approach reflects the tension between safety concerns and competitive pressures in AI development.

Path Forward

Stubb expressed hope that the declaration would create "momentum" and attract additional participants, including the two countries whose absence is most conspicuous. "I think it will be in the interest of both China and the United States to have some kind of guardrails," he said. "And our message here is that we need human intervention, so humans need to be in charge."

The initiative's success will likely depend on whether it can expand beyond its initial coalition to include nations that are home to leading AI companies and research programs.

These details were first reported by POLITICO.