What the panel is
The Independent International Scientific Panel on AI is the AI-governance equivalent of what the IPCC is for climate: a standing body meant to give the world’s governments a shared, credible reading of the science. It was created by the UN General Assembly in 2025 as one of the most concrete outcomes of the Global Digital Compact, and it moved fast: members appointed in February 2026, an inaugural meeting on 3 March, an in-person retreat in Madrid in April, a first briefing to UN member states on 19 June, and a preliminary report launched on 1 July alongside Secretary-General António Guterres.
Its defining feature is a deliberately narrow mandate. The panel is scientific, non-political, and policy-relevant but not policy-prescriptive: it documents where the evidence agrees and where it disagrees, and it explicitly does not make recommendations; that is the job of the intergovernmental Global Dialogue on AI Governance, to which the report was carried in Geneva in July. Its work was drafted by seven thematic working groups with a per-member sign-off; any member could have recorded a formal dissent, and none did.
Coverage
The preliminary report
“Evidence-based assessment of opportunities, risks and impacts of AI”: eight cross-cutting assessments and seven domains, in the panel’s first published output.
Read summary → 1 July 2026 · Press launchThe report goes public
Guterres says “we can no longer say we did not know”; Ressa and Bengio give three headlines and defend a report that documents but refuses to prescribe.
Read summary → 19 June 2026 · UNGA briefingThe first briefing to member states
Co-chairs Ressa and Bengio present the panel to the General Assembly: independence, the pace of AI, and the Global South’s central concern: the AI divide.
Read summary →What the first report says, in brief
The preliminary report is built around eight cross-cutting assessments: among them that AI capabilities are advancing faster than the ability to measure or govern them; that only a handful of actors have trained frontier models; that AI’s inputs and benefits are geographically and linguistically uneven; that the AI divide is about the capacity to shape AI, not just to access it; that agentic AI is a governance step change; and that AI can erode shared reality and is already reshaping human rights, including children’s. A fuller walk-through of the evidence is on the report summary page.
Why it matters for the SDGs
Every SDG now has an AI story: in health, education, agriculture, and the data systems that measure progress itself. The panel matters because it is the UN’s attempt to give 193 countries a common, non-partisan evidence base for those decisions, rather than leaving the science to be narrated by the handful of firms and states that build frontier models. Its sharpest SDG signal is the AI divide (SDG 10): more than 7,000 languages are spoken, but only about 1,000 have the foundations for meaningful AI inclusion, and fewer than a third of developing countries have a national AI strategy. And its “evidence dilemma” (that proof of AI’s impact lags the pace of its deployment) is a direct challenge to the measurement work at the heart of SDG accountability.
Key links
- Independent International Scientific Panel on AI, official UN panel page.
- Global Digital Compact, the framework that created the panel.
- UN Global Dialogue on AI Governance, where the report was presented (Geneva, 6–7 July 2026).
SDGCounting’s coverage draws on the panel’s published preliminary report and the record of its 19 June 2026 briefing to member states. Figures and findings are the panel’s own.