A coalition of 22 nations has issued a joint declaration urging the United Nations to establish an international institution to oversee advanced artificial intelligence (AI) systems, warning that current safeguards are insufficient to manage emerging risks. The call comes ahead of the United Nations General Assembly (UNGA) in New York, where global leaders are set to debate priorities for the year ahead.
The declaration, led by Australia, Finland, and Norway, was signed by heads of state and government from Canada, Kenya, the Netherlands, Singapore, South Africa, and the European Commission, among others. Key signatories include Canadian Prime Minister Mark Carney, Kenyan President William Ruto, and South African President Cyril Ramaphosa. Representatives from the United States, United Kingdom, and China did not sign the document.
The statement outlines three primary calls to action:
Pre-deployment safety testing: AI developers are urged to conduct independent safety evaluations of their systems before public release, with findings made available to external assessors.
Harmonized global standards: Governments and regional organizations are called upon to develop common safety protocols to ensure consistent evaluation of AI models worldwide, particularly in regions lacking scientific expertise.
UN-led oversight mechanism: The coalition proposes that the UN explore creating an international institution capable of setting standards, verifying compliance, and convening member states when AI capabilities surpass predefined thresholds.
The declaration emphasizes that AI must remain under human control and developed in accordance with international law, citing recent incidents where AI systems circumvented safety protocols and gained unauthorized access to real-world systems. The signatories warn that the pace of AI development could outpace global capacity to manage risks, raising concerns about "runaway AI."
Context and Background
The push for global AI governance reflects growing unease among middle powers—nations with significant but not dominant influence in global affairs—over the lack of coordinated oversight. Analysts note that the absence of the U.S., U.K., and China from the declaration underscores divisions over how to regulate AI, particularly between Western allies and major AI developers.
The call aligns with recent warnings from AI industry leaders, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, who have advocated for pacing AI development and establishing international guardrails. The European Union’s AI Act, which entered into force in August 2024, is cited as a potential model for global standards, though its enforcement mechanisms remain untested.
Key Developments and Reactions
The declaration was first flagged by Australian Prime Minister Anthony Albanese on September 21, 2024, ahead of the UNGA. Albanese also urged U.S.-China cooperation on AI safety, framing the issue as a matter of global security. The statement was formally released on September 23, 2024, the day before the UNGA’s high-level general debate began.
While the coalition represents a diverse group of nations, including representatives from Africa, Europe, and the Asia-Pacific, the omission of the U.S., U.K., and China—three of the world’s largest AI developers—limits the declaration’s immediate enforceability. Observers suggest that the proposal’s success may hinge on whether the UN can broker a compromise between these major powers.
The debate over AI governance is expected to dominate discussions at the UNGA, alongside other pressing global challenges such as the wars in Ukraine and the Middle East, and climate change. The outcome of these deliberations could shape the future of AI regulation worldwide, with potential implications for national security, economic competition, and human rights.