How can the voices of affected communities be placed central to AI governance? Across the world, citizens are sharing their views on the future of data, AI and technology through informed public deliberation.
This breakfast roundtable, organised to coincide with meetings of the UN Committee on Science and Technology for Development (CSTD) explored how transnational, national and grassroots public engagement can be brought into two-way connection with the evolving global AI Governance architecture.
Recording
A recording of the presentations from the session can be found here and embedded below. Slides are available here.
Discussion notes
Participants gathered to discuss the integration of a Citizens Track and wider deliberative and public participation processes into global AI governance frameworks, specifically focusing on the upcoming UN Global Dialogue on AI Governance, and the 2027 AI Summit in Switzerland. The discussion highlighted a critical tension between the fast-moving nature of AI development and the slower, often exclusionary, multilateral processes of global policy.
1. Entry Points in 2026 and 2027
- Organizational Status: Planning for the 2027 AI Summit is in early stages (targeted for Q2/Q3 2027). Organizers are committed to a multi-stakeholder approach, though the “citizen track” remains at the early concept stage, and not firmly on organisers agenda.
- Connecting to Mainstream Policy: It was noted that citizen deliberation is not yet embedded in current multilateral processes (e.g., the Global Digital Compact). To gain leverage, proponents must link citizen tracks to established bodies like the UN AI Scientific Panel, UNESCO, and the ITU.
- Public Information Barriers: While information about summits is technically public, it often fails to reach the “ordinary citizen.” Digital platforms (such as Swiss-hosted “Offcom” platforms) exist but primarily attract professional stakeholders rather than the general public.
2. Barriers to Inclusion & Energy to Engage
- The “Closed Door” Perception: Many citizens, local developers, and community leaders feel entirely excluded from policy spaces, viewing them as the domain of elite experts or big tech interests. We explored how there is high citizen interest to engage in dialogue around AI, but many are not sure where to get started.
- Fragmentation: There can appear to be a high degree of fragmentation across different summits (Geneva vs. New York) and initiatives. This makes it difficult for non-experts to know where or how to contribute effectively.
3. Strategic Recommendations for Engagement
- The “Offer” vs. The “Ask”: Rather than just demanding a seat at the table, how can citizen groups frame their participation as an “offer” of coordinated local action and unique insights that enhance the legitimacy of global decisions?
- Targeted Focus: It was suggested that participants use the existing language of policymakers (e.g., linking to the specific themes for the UN Global Dialogue) to make recommendations more digestible and actionable.
- Youth and Vulnerable Groups: There was a strong call to stop creating governance for the future without involving the youth who will live in it. Inclusion must also extend beyond “AI policy people” to include labor, health, and gender movements.
4. Innovation in Deliberative Tools & Approaches
- Decentralized Polling: Examples were shared of decentralized apps that allow for real-time, anonymous polling of public sentiment. Such technology could theoretically be “docked” into UN gatherings to provide immediate citizen feedback.
- Collective Intelligence: AI itself was proposed as a tool to help aggregate and synthesize vast amounts of public insight, potentially bridging the “speed gap” between technology and policy.
- Local Mobilization: Successful models from other fields (e.g., migration policy) show that creating formal “mechanisms” for mayors, youth, and civil society can eventually force states to recognize these voices as essential to the process.
Conclusion
The shift from multi-stakeholder models (like the Internet Governance Forum) to multilateral member-state-led processes was identified as a challenge to citizen inclusion. However, participants expressed cautious optimism that the 2027 summit offers a window to institutionalize citizen tracks. The immediate priority is to move from “talking about things” to running small-scale prototypes and ensuring that democratic institutions, rather than private firms alone, retain the power to shape the AI future.
The above notes were generated from the session transcript with AI assistance: cross-checked manually.
About
This session was organised by the Global Governance Centre, Geneva Graduate Institute, ISWE Foundation and Connected by Data with the Ada Lovelace Institute and partners.