The United Nations’ Global Dialogue on AI Governance is a test of whether the multilateral system can establish legitimate and inclusive AI oversight. Failure could accelerate regulatory fragmentation and make transboundary harm harder to manage.
Tony Oweke is a research fellow specializing in artificial intelligence (AI) governance and digital policy at the Council on Foreign Relations. Prior to joining CFR, Oweke represented a bloc of 134 developing countries in negotiations on AI at the United Nations.
The United Nations’ Global Dialogue on AI Governance—born out of the 2024 Global Digital Compact (GDC) and parallel UN negotiations that also produced an Independent International Scientific Panel on AI—will convene this July in Geneva. The Dialogue arrives amid a broader proliferation: a growing ecosystem of international AI summits, including in Bletchley Park, Seoul, Paris, and New Delhi, that are attempting to shape the global governance landscape. Whether these attempts are coalescing into something coherent, however, remains to be seen.
The 2023 Bletchley Declaration generated a shared international vocabulary around frontier-AI risk and established AI safety as a matter for coordinated global action. It also catalyzed the AI Safety Institute to publish an evaluation of advanced AI systems’ capabilities and risks in its annual report. The highlight of the 2024 summit in Seoul was a commitment to establish a network of AI safety institutes spanning ten countries and the European Union (EU). The 2025 summit in Paris marked a shift away from safety and toward investment and adoption, underlined by several funding pledges and deals. India’s 2026 summit, by contrast, sought to center voices from the Global South in AI governance. And early indications are that the 2027 summit in Switzerland will again focus on safety and security.
Taken together, these summits have succeeded in mobilizing political attention and resources, but they have struggled to translate high-level commitments into coordinated, durable governance outcomes.
In the wake of India’s February summit—which captivated global attention and featured an impressive array of participation from governments, big tech, and civil society—attention now shifts to the lesser-known, and almost certainly smaller, Global Dialogue on AI Governance. The question heading into the July summit is what the value proposition of this dialogue, and of pursuing AI governance through the United Nations, really is.
For many of the countries that drove its creation, that value rests in legitimacy. The United Nations, while restrained by geopolitical deadlock in the Security Council and funding shortfalls in development and humanitarian assistance, retains a convening power that compels countries to pursue global governance under its auspices—particularly in addressing and altering inequitable global structures. The logic is simple: the United Nations remains the sole universal platform through which every country can participate in shaping global governance.
That logic pervades several recent developments at the United Nations, including the pursuit of a Framework Convention on International Tax Cooperation to address the unfair allocation of taxing rights under the Organization for Economic Cooperation and Development’s two-pillar solution, as well as the establishment of a borrowers’ platform for indebted developing countries seeking to counter the largely creditor-driven international debt system. Many countries, burned by inequitable structures proving painful to correct, view pursuing AI governance through the United Nations as a preventive measure against future dependencies created by adopting technology and rules they did not help develop. For those countries, the Dialogue is a first and potentially consequential step toward more legitimate and inclusive AI governance.
Part of the Dialogue’s legitimacy lies in the possibility of producing concrete outcomes, a clear differentiator from the national summits. Although those summits have generated high-level political declarations, agreements, and pledges, the implementation and effects of those instruments remain uneven and unclear at best. Conversely, the summaries coming out of the Dialogues in Geneva and New York will feed into planned intergovernmental consultations designed to produce common priority areas for AI governance. Those consultations will, in turn, inform the review of the GDC, the document that established both the Scientific Panel and the Dialogue.
The GDC review—expected to occur sometime between October 2027 and August 2028—will provide an opportunity to take stock of those two institutions, assess whether they should be renewed, and, if member states deem it necessary, deepen the United Nations’ AI mandate. Doing so could include introducing a negotiated outcome to the Dialogue or launching an intergovernmental process aimed at producing an international legal instrument on AI. Conversely, member states could decide that these institutions are ineffective and shut down the organization’s AI mandate.
There is a broad range of possible outcomes, but a tangible thread exists between each and the Dialogue. In the interim, the dialogues will enable cross-stakeholder exchange of best practices to deepen global understanding of AI governance and influence the development of national policy for countries grappling with how to govern this transformative technology.
Should the United Nations fail to live up to such lofty aspirations, AI governance will continue to fall within the purview of exclusive multilateral clubs while incompatible national policies proliferate. Absent a credible central coordinating mechanism, fragmentation is likely to give way to normative competition, with states vying to diffuse governance models that lock in strategic, economic, and regulatory advantages as other countries adopt them. The discourse around competition currently revolves around first-mover countries and regions developing both AI systems and governance frameworks, such as the United States, China, and the EU.
But over the past two years, a new array of actors has begun jostling for influence over how AI will be governed transnationally in places like Delhi, Brasília, and Jakarta. As major diffusers and regional powers in their own spheres, those countries’ governments are increasingly seeking greater control over the capabilities, infrastructure, and governance arrangements that will shape their digital futures—and, importantly, influencing other countries to do so. With more than eighty countries advancing AI strategies or legislation based on different regulatory philosophies and developmental priorities, the resulting fragmentation increases the likelihood that AI deployed across borders will produce inconsistent outcomes, evade clear accountability, and amplify the risk of transboundary harm.
The United Nations’ Global Dialogue on AI Governance is not simply another series of meetings added to an already crowded calendar. It is a test of whether global efforts to govern AI can move beyond fragmented and incompatible national approaches and begin to take on a more coordinated and inclusive form. If it succeeds, it could help shape a more coherent international framework for managing the cross-border effects of AI. If it fails, governance will continue to splinter along geopolitical and economic lines, with rules increasingly set by those with the greatest capacity to extend their influence.
In that case, the central question will shift. It will no longer be whether AI is governed, but who governs it—and whose interests those rules ultimately serve. Much will depend on how these new institutions are designed, operationalized, and ultimately taken up by UN member states.
This work represents the views and opinions solely of the author. The Council on Foreign Relations is an independent, nonpartisan membership organization, think tank, and publisher, and takes no institutional positions on matters of policy.
This work is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) License.
By Tony Oweke
By Vinh Nguyen
©2026 Council on Foreign Relations
The World Is Trying to Govern AI. The UN Wants In. – Council on Foreign Relations
Leave a Comment
