The United Nations’ formal AI-governance track is no longer theoretical: the Global Dialogue on AI Governance held its first session on 6–7 July 2026 in Geneva, the Independent International Scientific Panel on AI released its Preliminary Report on 1 July 2026, and the Dialogue’s second session is already scheduled — 3–4 May 2027 in New York. Both bodies exist under a single UN General Assembly resolution adopted less than a year before that first session, and together they are the closest thing to a standing, multilateral AI-safety governance architecture the UN has produced.
What the 2025 resolution actually created
On 26 August 2025, the UN General Assembly adopted Resolution A/RES/79/325, establishing two interconnected mechanisms in a single instrument: the Independent International Scientific Panel on Artificial Intelligence and the Global Dialogue on Artificial Intelligence Governance. The resolution traces back to the Global Digital Compact (A/RES/79/1), adopted at the 2024 Summit of the Future, which committed UN member states to setting up an independent scientific panel so that international AI discussions would be, in the UN’s words, “informed by the best available evidence.” Costa Rica and Spain co-facilitated the negotiations that produced the final text, following a public process that ran from an initial January 2025 consultation through five successive revisions of the resolution’s draft text between May and August 2025.
Structurally, the two bodies split the work: the Scientific Panel is the evidence-generating half — independent experts producing assessments of how AI is transforming society — and the Global Dialogue is the negotiating half, a recurring intergovernmental forum where member states and stakeholders discuss what to do about it. Neither issues binding rules. What they produce is closer to the UN’s IPCC-for-AI model: shared, credentialed fact-finding paired with a standing venue for governments to talk to each other about it on a fixed cycle.
Key dates, so far
- 26 August 2025 — UNGA adopts Resolution A/RES/79/325, creating both bodies.
- April 2026 — The Scientific Panel holds its first in-person meeting, in Madrid, Spain.
- 1 July 2026 — The Scientific Panel releases its Preliminary Report, with a launch press conference carried on UN Web TV.
- 6–7 July 2026 — The Global Dialogue on AI Governance holds its first session, in Geneva.
- 3–4 May 2027 — The Global Dialogue’s second session is scheduled, in New York, alongside the 11th Multi-stakeholder Forum on Science, Technology and Innovation for the Sustainable Development Goals.
The Scientific Panel’s Preliminary Report
The Independent International Scientific Panel on AI released its Preliminary Report on 1 July 2026 — the same week as the Global Dialogue’s inaugural session, positioning the Panel’s evidence base directly ahead of the Dialogue’s first substantive conversation. UN Secretary-General António Guterres framed the Panel’s purpose plainly at its launch: “In a world where AI is racing ahead, this Panel will provide what’s been missing — rigorous, independent scientific insight.” The Panel had already convened its first in-person meeting two months earlier, in Madrid in April 2026, bringing together the international experts named to it. The UN describes the Panel as the first global scientific body dedicated specifically to assessing AI’s effects on society, modeled loosely on the evidence-first role the IPCC plays for climate policy.
The Global Dialogue’s first session: four thematic clusters
The Global Dialogue on AI Governance convened in Geneva on 6–7 July 2026, and its outcome is recorded in a Co-Chairs’ Summary from the Global Dialogue on AI Governance 2026, organized around four thematic clusters:
- AI opportunities and implications — the social, economic, ethical, cultural, linguistic and technical implications of AI.
- Bridging AI divides — capacity-building, access to AI applications and high-performance computing skills in developing countries, and the development of open-source software, open data and open AI models.
- Safe, secure and trustworthy AI — the development of safe, secure and trustworthy AI systems, and compatibility across different national and regional governance approaches.
- Human rights, transparency and accountability — respect for and protection and promotion of human rights, alongside transparency, accountability and robust human oversight of AI systems.
For a CASRAI audience, the third cluster is the one to watch most closely: it is the UN’s own attempt to reconcile the “compatibility across governance approaches” problem that boards, general counsel and AI-safety policy teams already live with day to day — the same jurisdiction-by-jurisdiction patchwork covered in CASRAI’s own AI Regulations Around the World guide. The Global Dialogue does not resolve that patchwork; it is the venue where the patchwork gets negotiated, on a now-fixed roughly annual cycle.
What’s scheduled next
The Global Dialogue’s second session is already on the calendar: 3–4 May 2027 in New York, timed to run alongside the UN’s 11th Multi-stakeholder Forum on Science, Technology and Innovation for the Sustainable Development Goals. That gives governance teams a genuine lead time to prepare rather than reacting after the fact — the kind of dated, trackable commitment that is otherwise rare in international AI governance, where most instruments are open-ended.
Why this is worth tracking now
Nothing here is binding, and neither body has enforcement power. What changed on 6–7 July 2026 is that the UN’s AI-governance track stopped being a single resolution on paper and became a body of accumulating, dated output: a Preliminary Report, a Co-Chairs’ Summary with four named thematic clusters, and a second session already on the calendar. For boards, general counsel and policy teams already tracking frontier-AI-safety obligations across national regimes, that is the difference between “the UN said it would do something about AI” and an actual, citable institutional record to check against — the same discipline CASRAI applies to lab-level commitments in guides like Commitment: Ten Labs and Regulators, One Inconsistent Promise and Seoul Frontier AI Safety Commitments, and to institutional coordination in The International AI Safety Institute Network. This UN track adds a distinct, higher layer above both: not a lab’s own pledge, and not a network of national safety institutes, but a standing intergovernmental forum with UN Secretariat backing and now a two-session track record.
Where NIKOLAI fits in
CASRAI’s own NIKOLAI project — an independent, unendorsed reference dictionary of 64 frontier-AI-safety elements across 10 tracks — does not yet have an element built specifically for multilateral institutions like the UN’s two new bodies; that is a genuine current gap in NIKOLAI’s coverage, not a claim CASRAI is dressing up as a fit. The closest existing element is N9’s Commitment, currently defined as “a public statement by an identified party (a developer, a government body, or a named individual signatory) to do or not do something, carrying an explicit modality (binding / unilateral / conditional / aspirational), a scope, an effective or target date, and a status value that can be tracked over time” — see NIKOLAI’s own Track System guide for how N9, Commitments & Governance, fits among NIKOLAI’s other nine tracks.
Resolution A/RES/79/325 itself maps onto that schema almost exactly as a worked example: the identified party is the UN General Assembly (a government-body-level actor under the Commitment element’s own definition), the modality is closer to binding-for-process than aspirational (it establishes standing bodies with a mandate, not just a statement of intent), the scope is global AI governance coordination, the effective date is 26 August 2025, and the status is now trackable across two dated milestones — the 1 July 2026 Preliminary Report and the 6–7 July 2026 Geneva session — with a third, 3–4 May 2027, already scheduled. That is precisely the kind of dated, status-trackable structure the Commitment element is built to hold. CASRAI has not added this specific UN architecture as a mapped row inside NIKOLAI itself; this article describes how it would fit, as an example, not a completed crosswalk.
FAQ
What is the UN’s Global Dialogue on AI Governance?
It is a recurring intergovernmental forum, established alongside the Independent International Scientific Panel on AI under UNGA Resolution A/RES/79/325 (adopted 26 August 2025), where UN member states and stakeholders discuss international AI governance. Its first session ran 6–7 July 2026 in Geneva; its second is scheduled for 3–4 May 2027 in New York.
What did the Independent International Scientific Panel on AI publish?
A Preliminary Report, released 1 July 2026 with a launch press conference on UN Web TV. The Panel, created by the same 2025 resolution, held its first in-person meeting in Madrid, Spain in April 2026.
What did the Global Dialogue’s first session actually cover?
Per the Co-Chairs’ Summary from the Global Dialogue on AI Governance 2026, four thematic clusters: AI opportunities and implications; bridging AI divides; safe, secure and trustworthy AI; and human rights, transparency and accountability.
Is any of this legally binding?
No. Both the Scientific Panel and the Global Dialogue are advisory and deliberative bodies under the UN General Assembly; neither issues binding rules on member states or AI developers. Their significance for governance teams is as a dated, trackable record of the multilateral conversation, not as a source of enforceable obligations.
Where can I read the primary sources?
Directly at the UN: the Global Digital Compact’s AI page (un.org/global-digital-compact/en/ai), the Global Dialogue on AI Governance site (un.org/global-dialogue-ai-governance/en), and the Independent International Scientific Panel on AI site (un.org/independent-international-scientific-panel-ai/en).
