The latter half of the 20th century was the “economist’s hour,” according to journalist Binyamin Appelbaum. Equipped with new population-level data and mathematical models, economists shaped policies by offering scientifically testable mechanisms for problem-solving and value creation. Today, as technological and societal disruptions upend established ways of working, it is time to ask: Whose hour is now upon us?
Many of the toughest challenges facing people and planet—such as those embedded in the world’s Sustainable Development Goals (SDGs)—are tough because they are problems that can only be solved together, through cooperation across diverse actors. Ending poverty and hunger or arresting environmental degradation globally demands a paradigm shift in capacities to work collectively within and across communities, sectors, and geographies.
For these issues, the current moment may belong not to a single discipline, such as economics, but to the transdisciplinary science of collective intelligence (CI).
Researchers across biological, behavioral, social, and computer sciences are leveraging advanced tools of the 21st century’s technological revolution—ubiquitous sensors, continuous data streams, and machine learning—to understand how collective behavior underpins life across scales, from multicellular organizations and ant colonies to human teams, organizations, and distributed digital networks. Representing a convergence of these approaches rooted in cognitive psychology, network theory, and computational modeling, CI provides a potentially unifying scientific framework for understanding how to sustain and steward societal and planetary systems.
In a similar way that economics shaped evidence-based policymaking in the 20th century, CI could serve as a scientific engine for identifying which collective approaches to the SDGs are most effective in specific contexts—and why.
Roundtable on CI for the SDGs
Motivated by this opportunity and building upon initial efforts to elevate CI for the SDGs, a September 2024 Brookings virtual roundtable convened over 40 leading CI researchers and sustainable development practitioners to explore how CI insights could enhance collective problem-solving efforts within and across sustainable development challenges.
In a lively and wide-ranging discussion, participants converged on three key domains where CI research and tools could help increase the effectiveness of collective approaches to sustainability: CI within multistakeholder teams and networks, CI for managing interdependencies between issues, and CI for scaling solutions across contexts.
Collective intelligence within multistakeholder teams and networks
Achieving any one SDG target, like mitigating the negative health impacts of extreme heat in urban areas (SDG 13.1), for example, requires collaboration among diverse stakeholders, including government agencies, civil society organizations, technology providers, and researchers—each holding a key to progress embedded with their own perspectives and professional jargon. Yet such collaborations often falter due to misaligned expectations, power imbalances, or ineffective group dynamics.
CI research offers actionable insights to improve collective performance in these settings. For example, a team’s “c-factor” (a statistical metric of a team’s intelligence, equivalent to “g-factor” capturing individual intelligence) is a powerful predictor of team performance across a range of laboratory experiments. Research has shown that a team’s c-factor, in turn, relies on team members developing mutually compatible expectations and abilities for shared action, known as “shared mental models” in cognitive psychology. Task complexity, collaboration process, group size, as well as cognitive capabilities such as social perceptiveness (or “Theory of Mind”) and physiological mechanisms (e.g., behavioral synchrony and physical exertion) combine in different ways to influence mental model alignment and team performance.
This evidence provides a foundation for investigating analogous mechanisms of CI in multistakeholder teams for specific SDG issues. For example, research could explore how training stakeholders to integrate diverse perspectives (e.g., combining indigenous knowledge with technical expertise) accelerates consensus-building or how equitable communication (using population-scale deliberation tools like Pol.Is, as demonstrated in the vTaiwan platform) can enhance SDG outcomes.
UNDP’s Accelerator Labs recently collaborated with CI experts and researchers to highlight and elevate CI-informed approaches to SDG challenges, such as UNDP’s Lao PDR Accelerator Lab use of group deliberation, ethnography, and micro surveys paired with satellite imagery to inform city-wide approaches to increasing recycling and composting (SDG 12).
Collective intelligence for managing interdependencies across issues
The SDGs are deeply interconnected: Progress on one goal can impact others. As such, efforts to alleviate poverty (SDG 1) need to be designed to not inadvertently compromise climate action through increased CO2 emissions (SDG 13) and vice versa.
In this context, roundtable participants considered how CI-driven tools could empower communities to navigate trade-offs between SDG issues and enable more approaches that advance multiple goals simultaneously. CI researchers proposed using computationalmodeling tools to quantify interdependencies between SDGs. SDG practitioners welcomed these approaches while also emphasizing that models must be grounded in local empirical data since SDG policy and implementation can vary widely across local, national, regional, or global contexts. For instance, in lowest-income countries, SDG-specific policies and infrastructures are often not yet mature enough for interdependencies between SDG actions to show up in current near-term empirical analyses.
Collective intelligence for scaling solutions
While progress on SDG issues hinges on locally led action tailored to cultural and ecological contexts, global systems are needed for distilling and sharing transferable knowledge and resources.
To this end, participants highlighted the potential of AI systems to help organize and scale effective collective solutions for the SDGs. For instance, emerging generative AI tools can support and aggregate group deliberations. However, participants stressed that AI systems must be designed to support human agency and action at all levels of social organization, while preserving important socio-emotional underpinnings of CI, including trust, belonging, solidarity, and collective will.
Beyond AI, participants identified universities as natural hubs for testing and applying CI approaches, given their role as centers for transdisciplinary knowledge and cross-sector collaboration. Other innovative models for scaling CI include globally representative citizens assemblies, locally rooted “impact hubs” promoting partnerships for targeted outcomes (such as those related to maternal health), and experimental approaches like Operation Warp Speed that seek to rapidly align diverse stakeholder incentives and action. Participants also called for global mechanisms to consolidate evidence-based recommendations on collective behavior—similar to the IPCC’s role in climate science—to guide global sustainable development in an era of emerging technologies like AI.
Next steps: Developing a shared approach to applying CI to sustainable development
Participants identified three practical actions that could be advanced in the next 12-18 months:
Prototyping a shared research protocol. Participants identified an opportunity to use generative AI tools to efficiently collate and analyze case studies of existing collective solutions for the SDGs—developed across initiatives like People First Community, Reach Alliance, SDG Fund, Solutions Insights Lab, Sustainable Development Solutions Network, and the UNDP Accelerator Labs’ solutions mapping project, among others—to surface, test, and refine CI mechanisms tailored to specific SDG solutions. This effort could include organizing CI approaches to SDG issues across dimensions of action type, geography, scale, and maturity, and developing common metrics like a “c-factor” for multistakeholder collaboration in teams, networks, and ecosystems.
Establishing common research foundations. A special journal edition of Collective Intelligence, dedicated to elevating new and existing efforts to apply CI to sustainable development outcomes, could deepen the academic and practitioner community’s shared understanding of CI’s role in SDG action and identify practical initiatives to bridge CI research and sustainable development practice.
Growing collaborative networks of CI researchers and development practitioners. Building on an initial pilot conducted at the 2024 ACM Collective Intelligence conference, a full-scale 17 Rooms-style convening among CI researchers, SDG practitioners, and AI experts could foster high-value collaborations and equip practitioners with evidence-based CI tools.
Conclusion
In a time of disruptive change, the science of CI can help tackle challenges that can only be solved together. By developing a shared protocol for surfacing, testing, and refining CI mechanisms for the SDGs, scientists and practitioners can create evidence-based tools for multistakeholder collaboration, managing interdependencies, and scaling context-calibrated solutions for people and planet.
Roundtable participants
Disraeli Asante-Darko, Ashesi University
Joseph Bak-Coleman, University of Konstanz
David Baltaze, Unanimous.ai
Ruairidh Battleday, Harvard University
Joshua Becker, University College London
Javier Antonio Brolo, UNDP
Han Sheng Chia, USAID
Pablo Diego-Rosell, Gallup
Steven Dow, University of California San Diego
Carlos Estuardo Mazariegos Orellana, UNDP
Stefanie Falconi, USAID
Dan Foy, Gallup
Olamide Goriola, ConvergenceAI
Julian Gullet, Northeastern University
Kippy Joseph, Results for Development
Rafael Kauffman, Digital Gaia
Tom Kehler, CrowdSmart.ai
Drew Keller, Harvard University
Moni Kim, University of Toronto
Thomas Kopinski, University of South Westphalia
Gordon LaForge, New America
Enoch Li, INSEAD
Marin MacLeod, The Reach Alliance
Camille Masselot, Learning Planet Institute
John McArthur, Brookings Institution
Nicholas Mohnack, BundleIQ
Scott Page, University of Michigan
Chris Pease, MIT
Niccolo Pescetelli, Psi.tech
Christoph Riedl, Northeastern University
Louis Rosenberg, Unanimous.ai
Radha Ruparell, Teach For All
Adam Russell, University of Southern California
Iza Sanchez Siller, Tec de Monterrey
Saiph Savage, Northeastern University
Daniel Shusset, Villanova University
Dilip Soman, University of Toronto
Georg Theiner, Villanova University
Bryce Tully, Innerlogic
Fergus Turner, University of Cape Town
João Paulo Vergueiro, Giving Tuesday
Joe Wong, University of Toronto