A coalition of 20 nations, led by Norway and Finland, has called for the United Nations to create an international body to oversee the development and use of frontier AI models. The declaration, adopted ahead of this week’s U.N. General Assembly in New York City, urges the U.N. to explore establishing “an international institution, able to set standards, enable verification, and convene states when capability thresholds are crossed.”
The push comes as several major technology leaders have issued warnings about AI. The declaration also echoes a proposal recently pitched by Anthropic CEO Dario Amodei.
The coalition’s demand
The declaration was signed by 20 nations, including Germany and Canada, as well as the European Commission. It is titled “A call for control of frontier AI models.”
The document’s core recommendation is that the U.N. explore creating an international institution to set standards, enable verification, and bring states together when AI capability thresholds are crossed. The recommendation is presented as a response to the growing power of AI systems.
The declaration also contains a separate recommendation that urges governments and regional organizations to develop and coordinate “common standards” to ensure that countries across all regions have access to scientific capacity, expertise, and trusted evaluation. This recommendation appears to target the issue of technological imbalance across nations.
The push gained momentum as the U.N. General Assembly opened in New York City.
The resemblance to Amodei’s proposal
The declaration’s recommendation bears some resemblance to a global structure that Anthropic CEO Dario Amodei recently pitched. The coalition’s call for an international institution to set standards and verify capabilities aligns with the direction Amodei has pointed.
The comparison is notable because it puts a single company’s proposal at the center of a multilateral diplomatic push. The coalition is not simply endorsing Amodei’s work; it is citing it as the starting point for its own demand.
“an international institution, able to set standards, enable verification, and convene states when capability thresholds are crossed”
What the declaration asks for
The declaration’s central recommendation is that the U.N. explore creating an international institution. The institution would have three core functions:
- Set standards for the development and use of frontier AI models
- Enable verification of those standards
- Convene states when AI capability thresholds are crossed
The threshold language is significant. It suggests the coalition is looking for an institution that would not just observe AI development but would also trigger action when systems reach certain levels of capability.
The declaration also calls on governments and regional organizations to develop and coordinate common standards. This recommendation is aimed at ensuring that scientific capacity, expertise, and trusted evaluation are available across all regions.
The coalition’s framing suggests a desire to avoid a future where only a few wealthy nations have access to the most advanced AI systems. The declaration’s authors want to ensure that the technology’s benefits are distributed evenly.
The timing of the push
The coalition’s declaration comes as the U.N. General Assembly opened in New York City. The General Assembly is the annual gathering of world leaders.
The timing is deliberate. By presenting the declaration at the start of the assembly, the coalition is putting the issue of AI oversight on the global diplomatic agenda. The push also follows a wave of warnings from technology leaders.
What the coalition is not asking for
The coalition’s declaration is not asking for a single global regulator with enforcement power. It is asking the U.N. to explore creating an institution that would set standards, verify compliance, and convene states when thresholds are crossed.
The distinction matters. A regulator with enforcement power would require a legal framework that the U.N. does not currently have. The coalition’s proposal is more modest: it is asking for an institution that would convene states and bring them together to set standards.
The verification function is also important. It suggests the coalition is not just interested in setting standards but also in checking that they are being followed.
Why the coalition matters
The coalition’s membership is significant. Norway and Finland led the declaration, and the 20 signatories include Germany and Canada, as well as the European Commission.
The coalition’s approach is multilateral. It is asking the U.N. to explore an institution, rather than creating one unilaterally. That means the coalition is working within the existing framework of international diplomacy.
The coalition is also pragmatic. It is not calling for a complete ban on AI or a moratorium. It is asking for standards, verification, and convening power.
The comparison to Amodei’s proposal is also notable. The coalition is not simply endorsing a private company’s work; it is citing it as the starting point for its own demand. That suggests the proposal has gained real traction in diplomatic circles.
The coalition’s next steps
The coalition’s declaration is a call for action. The U.N. is expected to explore the proposal, and the coalition will likely push for concrete steps to be taken.
The coalition’s membership gives it influence. The 20 signatories represent a range of regions and interests.
The declaration’s authors will also be watching how the U.N. responds. If the U.N. moves forward with exploratory work, the coalition could become a driving force in shaping the institution’s design.
The coalition’s call for common standards is also significant. It suggests the coalition is interested in a shared global framework, rather than a patchwork of national regulations.
The coalition’s declaration is a call for action. The U.N. is expected to explore the proposal, and the coalition will likely push for concrete steps to be taken. The coalition’s membership gives it influence, and its pragmatic approach makes it harder to ignore.
Where the paper stands
The paper backs narrow, targeted regulation against direct harm by AI companies, such as forcing disclosure of safety failures they hid, and is against broad international oversight and licensing that would freeze today’s leaders in place and lock out whoever would have challenged them. The paper opposes rules that raise the cost of entry for small firms, and this coalition’s push toward a U.N. institution with enforcement-like powers risks doing just that.
The coalition cites a single company’s proposal as the starting point for a multilateral diplomatic push. A U.N. institution with standards-setting and verification powers could easily become a moat, not a safeguard — a licensing regime and compliance cost that only the biggest firms can afford, and that freezes today’s leaders in place. The paper’s position is that regulation should target harm directly, not build an apparatus that hands the market to the incumbents.
The coalition’s next steps will show whether it is pushing for a real institution or a fig leaf. The paper will watch how the U.N. responds, and whether the coalition’s pragmatic framing holds or hardens into a regime that locks out whoever would have challenged today’s leaders.
See the a run of 14 images at the Washington Examiner.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.