The United Nations University Centre for Policy Research (UNU-CPR) and the British Academy convened roughly a day-long workshop in New York on 29 October, dedicated to redesigning science-policy interfaces (SPIs) — the institutional machinery that turns scientific assessment into multilateral decision-making.

The meeting sat inside UNU-CPR's Global Governance Innovation Platform (GGI), a programme exploring governance structures and decision-making mechanisms. A working paper synthesizing the workshop's insights is in development and is intended to anchor further cross-disciplinary collaboration.

Fifty years of institutional accretion

Participants traced the proliferation of SPIs back to the 1972 Stockholm Conference on the Human Environment. Since then, the Intergovernmental Panel on Climate Change has become the reference model, but new interfaces now operate across biodiversity, chemicals, health, oceans and artificial intelligence — each shaped by the constraints of its own policy environment and knowledge base.

The high-level panel drew representatives from the UN Secretary-General's Executive Office, the World Bank, the World Meteorological Organization, IPBES, and Professor Tshilidzi Marwala, Rector of the United Nations University and Under-Secretary-General. That composition signals who the workshop considered the actual customers of SPI reform: institutions that buy, synthesize and act on scientific evidence at scale.

Credibility as a design problem

Speakers treated credibility not as a by-product of technical quality but as an institutional design outcome. Three design features recurred: independence, inclusivity and clarity of mandate. Drawing on the experience of the UN Secretary-General's Scientific Advisory Board, participants argued that credibility depends on the diversity of perspectives and proximity to decision-making processes, not expertise alone.

The Global AI Panel served as the second case study — an expert-led, independent model with a mandate to deliver evidence-based advice that is transparent and globally representative. For R&D managers watching AI governance evolve, the panel is the closest analogue to what an intergovernmental advisory body for their field might look like.

Where the models strain

The workshop did not shy away from known failure modes. Participants acknowledged that impact pathways from assessment to policy are rarely linear, and that contestation and negotiation are integral to knowledge uptake — a concession that SPI outputs cannot be evaluated like journal citations. Two structural fixes emerged: clear feedback loops between assessment and decision-making, and sustainable financing to keep institutions effective over the long term. Neither is guaranteed in current UN budgeting practice.

Access inequalities also drew attention. Equitable access to information remains uneven across countries, and participants argued international institutions must do more to serve as credible, reliable data sources for policymakers in resource-constrained environments. Calls for stronger open science frameworks, particularly to widen access for the Global South, followed the same logic.

Discussions on knowledge integration highlighted the value of procedural and epistemological equity when incorporating Indigenous and local knowledge — an operational challenge for any assessment body built around peer-reviewed literature. Participants also expressed interest in more systematic collaboration among SPIs to address interconnected challenges underpinning the Sustainable Development Goals.

Health and oceans as testbeds

Two sectoral experiments received specific scrutiny. In health governance, participants examined efforts to build standing mechanisms for synthesizing evidence on antimicrobial resistance and pandemic preparedness — effectively a move from ad hoc review panels toward continuous evidence infrastructure. In environmental governance, the World Ocean Assessment provided the case study on integrating scientific input into intergovernmental processes.

The health models are worth tracking for research organizations: standing synthesis mechanisms, if adopted, would create recurring demand for evidence review rather than one-off commissioned reports.

Lag time as the core metric

Underneath the institutional vocabulary, the workshop's operating concern was latency — the lag between scientific discovery and policy response. Participants argued that shortening it requires two parallel investments: stronger dialogue between scientists and policymakers, and building scientific literacy inside policymaking institutions themselves. Science, several speakers noted, functions as a cornerstone of multilateral cooperation and a global public good.

The closing discussion returned to first principles: differences across scientific fields shape both institutional design and the credibility of evidence in global governance. SPIs, participants concluded, must balance scientific independence with political utility while sustaining cross-disciplinary dialogue and continued collaboration between science, policy and the UN system.

The GGI working paper now in development will determine whether the workshop's design principles — independence, inclusivity, clear mandates, feedback loops, sustainable financing — translate into actionable institutional recommendations or remain conference-level consensus.