What the UN Global Dialogue on AI Governance Reveals About Global Power Shifts

Critical Questions by Laura Caroli and Matt Mande

Published October 7, 2025

On September 25, the United Nations launched the Global Dialogue on AI Governance. The dialogue aims to provide a platform for future discussions of AI governance. Governments and other stakeholders will convene annually-starting at the 2026 AI for Good Global Summit in Geneva-to discuss the safe development of AI systems, AI capacity gaps in developing countries, interoperability of national AI governance efforts, and socioeconomic implications of AI technologies.

The Global Dialogue on AI Governance is a milestone in the quest for multilateral AI governance, but not all countries are in support. At a UN Security Council debate the day before the launch, the United States came out in strong opposition to any and all multilateral AI governance initiatives, casting doubt on the future of the dialogue and on its meaningfulness.

Q1: Where does the Global Dialogue on AI Governance stem from?

A1: This launch was a year in the making. In a report published in September 2024, the UN High-Level Advisory Body on AI recommended the creation of a scientific panel on AI, akin to existing initiatives to assess risks of climate change and atomic radiation. The report also proposed a recurring policy dialogue that would draw on the panel's technical expertise and work towards a multilateral AI governance framework. At the Summit of the Future the same month, the United Nations committed to establishing the Independent International Scientific Panel on AI and the Global Dialogue on AI Governance in the Global Digital Compact. After many months of negotiations led by Spain and Costa Rica, the General Assembly finally approved the joint initiatives by consensus in an August resolution. In the words of Secretary General António Guterres,

The goals of the Global Dialogue are clear:

To help build safe, secure and trustworthy AI systems-grounded in international law, human rights and effective oversight; To promote interoperability between governance regimes-aligning rules, reducing barriers and boosting economic cooperation; And to encourage open innovation-including open-source tools and shared resources-accessible to all.

Guterres continued by highlighting the added value of the Global Dialogue: "For the first time, every country will have a seat at the table of AI." He also explained that "it will complement existing efforts around the world," citing the work of the Organisation for Economic Co-operation and Development (OECD), G7, and other regional efforts.

Q2: What did speakers say about AI at the United Nations?

A2: Countries were mostly united in calling for cooperation on multilateral AI governance, particularly to address the "digital divide" between the Global North and South, albeit with national nuances.

Rather than treating the dialogue as a low-cost diplomatic gesture, the United States was alone in pushing back. Office of Science and Technology Policy Director Michael Kratsios explicitly rejected "centralized control and global governance" of AI, signaling skepticism not only of UN-anchored rule-setting but also of voluntary compacts.

In contrast, China expressed strong support for a global governance framework. It aligned itself closely with developing countries, pushing for all nations to have a seat at the table lest AI governance become "a game of the club of wealthy nations." Iraq echoed this point when speaking on behalf of G77 and China, emphasizing that the Global Dialogue on AI Governance should take meaningful steps to address disparities in AI infrastructure, technical capacities, and technical skills.

Though the United States was absent from the dialogue launch, U.S. tech giants Microsoft and Meta both joined the discussion. Microsoft President Brad Smith repeated calls to bridge capacity gaps between nations, highlighting the role of private sector investment in building out key AI infrastructure. Meta similarly pushed for public-private collaboration on AI, but signaled solidarity with the White House and encouraged other countries to follow the example of America's AI Action Plan.

Q3: How did countries' statements compare to previous international conversations on AI?

A3: Most countries echoed positions consistent with prior multilateral initiatives on AI, but the United States' opposing statement at the UN Security Council marked a sharper break. The United States took a similar posture earlier in 2025 when it declined to sign the nonbinding Paris AI summit declaration, though officials had not explicitly rejected global governance at that time. This stance contrasts with U.S. positions in prior dialogues under both Trump's first term and the Biden administration. In 2019, Washington endorsed the OECD AI Principles and welcomed the G20's nonbinding, "human-centered" AI principles; in 2020, it helped found the Global Partnership on AI. Biden-era diplomacy then layered on the Bletchley Declaration in 2023 and a U.S.-led UN General Assembly AI resolution, along with the Council of Europe AI treaty, in 2024. Together, those steps embraced the role of international institutions in governing AI, an approach that the second Trump administration seems to have departed from.

By contrast, China's stance at the United Nations was broadly coherent with past positioning: elevating international bodies to build inclusive, consensus-driven governance. China took the same line at Bletchley, during UN General Assembly work on AI capacity-building, and at the Paris AI Action Summit, which China signed.

Q4: Why does the Trump administration reject global AI governance?

A4: The Trump administration views global AI governance by an international body as an impediment to innovation and the diffusion of U.S. values, key strategic goals in the administration's AI race against China. This belief was made clear in the administration's AI Action Plan, released in July, which stated that many international governance efforts tend to advocate for "burdensome regulations" and "cultural agendas that do not align with American values," and further are subject to the increasing influence of the Chinese Communist Party (CCP).

Instead, the administration prioritizes setting de facto standards and governance through the diffusion of a U.S. technology stack. The July executive order on AI exports frames its objective as global adoption of U.S. technologies, standards, and governance models to secure continued U.S. technological dominance. Office of Science and Technology Policy Director Michael Kratsios emphasized this approach during the UN Security Council's discussion on AI, describing the administration's ambition to make American AI the global benchmark and to help partners build "sovereign AI ecosystems" on secure U.S. infrastructure.

The administration's stance on global governance also reflects its skepticism of existential, large-scale risks from AI that would merit such rulemaking. In his remarks, Kratsios described existential risk as an "ideological fixation," along with social equity and climate "catastrophism," despite other nations repeatedly expressing concern about these risks in their statements.

The AI Action Plan lists "Counter Chinese Influence in International Governance Bodies" as a key policy action. Rather than achieving this policy through leadership in international bodies like the United Nations, Kratsios' remarks at the UN Security Council show a resolve to minimize the role of international bodies altogether.

Q5: Why is China in favor of global governance efforts?

A5: The CCP views UN-based global governance as an opportunity to blunt U.S. leadership in AI and cultivate soft power, especially as the U.S. withdraws from international conversations.

During the UN Security Council discussion, Vice Foreign Minister Ma Zhaoxu argued against "unilateralism and protectionism," instead supporting a consensus-driven approach. In his remarks, Ma condemned "high fences around a small yard" and "artificially imposed technological barriers," clearly referencing U.S. export controls and related measures. He went on to call for norms, standards, and governance models developed through international consensus, an approach that would dilute U.S. influence.

Ma's remarks track with the CCP's broader push to gain influence in international venues and cultivate a positive image. China's Global AI Governance Action Plan, published a mere three days after the U.S. AI Action Plan in July, casts the United Nations as the main stage for AI governance and links rulemaking to "bridging the digital divide" and building an "inclusive and fair" system. In parallel, China has expanded its footprint across international standards bodies and invested heavily in standards work in the past few years, albeit with mixed results. This institutional approach, paired with rhetoric meant to project a "credible, loveable and respectable China," underscores China's support for centralized governance.

Ultimately, the United States and China both compete for global leadership on AI. Their approaches diverge strategically and rhetorically, with the United States favoring de facto dominance through diffusion of its technology and China opting to shape AI's trajectory through global institutions.

Q6: What do the conversations at the United Nations mean for global AI governance?

A6: Washington's rejection substantially reduces the dialogue's near-term relevance, since meaningful global governance is difficult without the participation of the leading technology holder. It also reinforces the leadership vacuum that China appears ready to fill, both within AI discussions and across UN forums more broadly. These elements, paired with the difficulty of reaching consensus among so many different national perspectives, raise questions about whether the United Nations is the ideal seat for substantial action around AI governance, although the broad participation and the convening of meaningful conversations are equally important achievements at this stage.

Indeed, beyond the U.S. and Chinese positions, the broader conversations around the Global Dialogue should not be overlooked, as they reveal important dynamics for international AI governance. Like many UN resolutions, the text establishing it reflects compromise, and the "minimum common denominator" outcome underscores persistent disagreements on substantive governance questions.

Global South participation is one of the dialogue's clearest signals, cementing a trend of growing engagement in recent months. In their speeches at the dialogue and during the vote on the resolution that established it, many governments emphasized sovereignty in AI governance and the prerogative of states to shape the future Scientific Panel. This reflects an ambition to ensure developing countries are not merely norm-takers but active partners, though it also raises questions about how non-democratic states may influence the process. The exclusion of military uses of AI further illustrates where states prefer to keep discussions within separate, security-focused forums.

Moreover, Western states have generally supported the initiative, motivated both by reputational incentives and by opportunities to align the dialogue with broader goals of inclusivity and responsible innovation. However, they diverge from Global South positions on the role of expertise: Many Western states favor an expert-driven, merit-based, and independent process, in tension with sovereignty-first preferences.

The private sector dimension should also not be overlooked. Engagement by major technology firms-especially U.S.-based companies-suggests that while Washington remains skeptical, these companies see value in the dialogue's focus on building trust in AI and bridging the digital divide. This also helps with promoting their own "AI for Good" initiatives and boosting their reputation, in addition to opening further business opportunities. The U.S.-based companies' involvement indeed points to an alternative layer of legitimacy and influence beyond intergovernmental negotiations-one they are well-positioned to occupy, as gatekeepers of the technology, even if this unsettles some Global South governments.

In sum, the UN Global Dialogue on AI Governance matters less for its immediate regulatory impact than as a powerful symbolic signal of political will. As already highlighted by the leaders' declaration at the Paris AI Action Summit, it reflects a shift in global conversations toward ensuring AI benefits the many rather than the few. The Global South has emerged as a vocal stakeholder in this process, with China amplifying its leadership role. The forthcoming India AI Impact Summit will likely reinforce this trajectory and deserves close attention.

The UN Global Dialogue on AI Governance also functions as a testing ground for deeper tensions. How the Scientific Panel is constituted will reveal whether states or independent experts hold the balance of influence. Meanwhile, U.S. disengagement and China's growing assertiveness make the dialogue an important barometer of shifting geopolitical dynamics in AI governance.

Laura Caroli is the senior fellow of the Wadhwani AI Center at the Center for Strategic and International Studies (CSIS) in Washington, D.C. Matt Mande is a research assistant with the Wadhwani AI Center at CSIS.

The authors wish to thank Nicole Errera for her research support for this article.

Programs & Projects