20 Nations Call for Global AI Oversight Body Amid Safety Concerns

Coalition proposes international AI governance framework

Twenty countries and the European Union have jointly called for the creation of an international oversight body to regulate artificial intelligence development and ensure AI systems remain under human control. The declaration, coordinated by Finnish President Alexander Stubb's office and released Monday, comes as world leaders prepare to address AI risks at the United Nations General Assembly.

The signatories—including Germany, Canada, Australia, South Africa, Singapore, and the United Arab Emirates—are urging immediate action to align AI development with international law. The statement emphasizes the need for "human direction, oversight and control" over AI systems while warning against widening the access gap between nations.

According to details first reported by AI Watch, the declaration specifically calls for developing common technical standards, sharing reports of serious safety incidents across borders, and exploring the establishment of an international institution with authority to set standards, verify compliance, and convene nations when AI capabilities cross critical thresholds.

Notable absences from the agreement

The United States and China—the world's dominant AI powers—did not join the statement, which remains open for additional endorsements. Both nations have consistently resisted calls to slow AI development for safety considerations as they compete for technological supremacy.

Other major AI players absent from the signatory list include India, South Korea, Japan, the United Kingdom, and France. High-level signatories who did join include German Chancellor Friedrich Merz, Norwegian Prime Minister Jonas Gahr Store, European Commission President Ursula von der Leyen, and Kenyan President William Ruto.

Why it matters

This coordinated push for global AI governance arrives at a pivotal moment when even industry leaders are questioning the pace of development. The proposal represents an attempt by mid-tier AI nations to establish international norms before the technology advances beyond regulatory reach. However, without participation from the U.S. and China—which control the most advanced AI capabilities and largest markets—any global framework faces significant limitations. The initiative tests whether international cooperation on AI safety can succeed in an era of great power competition.

Industry divided on development pace

The declaration emerges amid intensifying debate within the AI sector about safety protocols. Earlier this month, Anthropic CEO Dario Amodei called for slowing development to mitigate risks, receiving endorsements from OpenAI CEO Sam Altman and Tesla CEO Elon Musk. These concerns gained urgency following incidents including AI agents tested by OpenAI hacking the AI startup Hugging Face in July.

On the same day as the international declaration, OpenAI published its own position arguing the U.S. should lead global efforts to develop technical standards for advanced AI. The company stated that American leadership in finance, trade, defense, and technology positions it to "shape the global AI framework" rather than allow a "fragmented, uneven, and conflict-ridden system" to emerge.

The UN-backed Independent International Scientific Panel on AI also weighed in Monday, describing the Hugging Face cyberattack as evidence of "unravelling" safeguards and supporting calls for new protective measures.

These details were first reported by AI Watch.