The first UN Global Dialogue on AI Governance will take place in July 2026.
The co-chairs invited written submissions to help shape the dialogue, and held a number of virtual and in-person consultations.
We prepared a PAIRS and Citizens Track submission to the written consultation.
The consultation response draws upon:
- Informal conversations at PAIRS and with Stewardship group
- Citizens Track Webinar - 23rd March
- Geneva Roundtable - 23rd April
- Online feedback
Our responses to the key questions are listed below:
8. In your opinion, what outcomes would make the first Global Dialogue on AI Governance a success?
The UN Global Dialogue on AI Governance can foster an inclusive global conversation on AI: rooted in our common good. To deliver this, it should leverage the appetite of publics across the world to be included in debate around AI: supporting depolarisation and strengthening mandates for coordinated governance action delivering both minimum expectations of AI systems and shared visions for AI’s role in the future.
The first AI Dialogue can lay the groundwork for a ‘Citizens’ Track’ that supports this by:
-
Integrating public voices on AI: including asking the Scientific Panel report to address evidence gaps on public attitudes, and making findings from existing public dialogues on AI visible across the Dialogue venue and sessions, while supporting civil society participants to highlight voices from affected communities. A successful dialogue is one anchored in everyday citizen concerns, not only expert, industry or elite input.
-
Catalysing inclusive public deliberation on Dialogue themes and outcomes: identifying key questions that need further widespread public deliberation, and inviting creation of decentralized citizen assemblies across the world that can learn about and discuss these issues, feeding back to the 2027 Global Dialogue. A successful dialogue communicates by inviting communities to build their agency and engagement with AI governance, rather than projecting distance, complexity or intractability.
-
Embedding public participation into any commitment and review mechanism. The true test of any governance action is whether the community supposed to benefit can see that benefit. Voices of affected communities must be part of any commitment audit or monitoring mechanism.
-
Ensuring equal and meaningful civil society participation including addressing structural barriers to majority world civil society participation such as funding, visas, advanced notice of agendas, and equal access to networking spaces.
A successful AI Dialogue will establish the mandate to institutionalise public participation within future iterations of the meeting.
From your perspective, which of the following thematic areas identified by the General Assembly Resolution 79/325 for the AI Dialogue reflect your priorities for urgent action and active engagement by your entity? (Please select up to 4 priorities)
- AI capacity-building
- Social, economic, ethical, cultural, linguistic and technical implications of AI
- Interoperability of governance approaches
- Transparency, accountability, and human oversight
10. Please briefly explain your selection
A standing Citizens’ Track on AI governance, as a part of the ‘Dialogue of dialogues’, has potential to support interoperability of governance approaches, and mobilise public support for governance action at all levels: global to grassroots.
Evidence from existing public deliberation suggests, when given space for informed discussion, citizens across the world identify many common hopes and fears about AI, and common desire for empowerment shaping AI in their lives.
Anchoring narratives of global AI governance in shared global citizen concerns, and linking local action and global governance, can address polarisation of AI governance debates, building support for collective action that directs development of AI towards sustainable public good: realising potential and addressing harms.
We place related priority on AI capacity-building and emphasise building critical AI literacy amongst broad populations in order to support inclusive and participatory democratic shaping of AI development, deployment and governance.
We note the need for a broad and inclusive debate about capacity-building, not solely focussed on technology access and AI skills, but also addressing capability of citizens to make active choices about if, how and when to integrate AI into daily life, to explore different sustainable futures alongside AI, and to engage with reporting and redress mechanisms.
Inclusive and pluralistic capacity building is critical to allow different geographic, linguistic and cultural communities to define their own democratic paths with AI: as opposed to assuming only one path towards AI adoption.
We prioritise transparency, accountability and human oversight as critical components of the democratic governance of AI. We note that human oversight should cover design and pre-deployment decisions, as well as oversight of AI in action.
Our work on participatory governance emphasises the necessity of grounding discussions of AI in real-world lived experience of AI impacts: addressing social, economic, ethical, cultural and linguistic implications of AI.
11. In your opinion, are there any cross-cutting or emerging issues not captured by the listed themes above? If so, please explain.
The democratic governance of AI is a critical cross cutting issue. Any global AI governance frameworks need to consider principles of subsidiarity, and whether they are enabling relevant decisions to be taken by affected communities at the lowest possible level, taking into account universal human rights and interoperability of governance approaches.
Across all themes it will be important to ask about current public attitudes towards the issues and actions being explored. This will ensure that expert inputs are interpreted alongside public priorities, and in line with expressed public interests.
Human rights is the framework within which the other thematic areas operate. As convenor, the UN brings its Charter and Mandate including its role as the custodian of the international human rights system. Human rights are a cross-cutting foundation.
12. How are the governance gaps and related developments/advances in the thematic areas you selected above affecting your country, region, or sector? Please highlight the most significant challenges and opportunities.
There is substantial public demand for action on AI governance. Global survey evidence (2025 KPMG global survey of public attitudes towards AI, figure 24) [1] shows that across 42 countries, the majority of people in all countries (70%) view AI regulation as a necessity. This may be increasing: the Ada Lovelace Institute / Alan Turing Institute UK public attitudes survey shows 72% of people surveyed in 2025 want AI legislation, rising from 62% in 2023, in a regulatory environment where legislation has not been forthcoming [2].
In addition, we see particular gaps with respect to independent capacity building efforts - with a need to ensure capacity building takes place in the public interest without creating vendor lock in and dependence. As per the 2025 UNDP Human Development Report, effective capacity building must contribute to increased human agency with AI, rather than introducing tools that limit individual and collective agency.
With respect to transparency, accountability and oversight, papers at the Participatory AI Research & Practice Symposium (PAIRS) [3] have explored the importance of both participatory elements to official AI audits, and community-led audits of AI systems. These both benefit from clear transparency frameworks, and recognition of the role of affected communities within audit.
[2]: https://attitudestoai.uk/
13. What role can the AI Dialogue play in advancing international cooperation on AI governance?
AI governance involves decisions and action at all levels, from the individual workplace, classroom, forest or field, to the actions of regional, national and global firms, governance institutions and civil society stakeholders. The Dialogue cannot be the container for all these conversations, but it can provide an anchor point for many of them - and it can identify key gaps, where important voices are missing from the conversation, or where key questions need deeper public deliberation. It can elevate global majority voices and leadership on critical issues of AI governance, and can foster new coalitions testing approaches to people-centred AI governance.
The Dialogue can advance international cooperation on a ‘governance floor’ for AI, building a mandate for the required governance actions that can only be undertaken at a global level, providing the foundation for conversation at other levels whilst supporting interoperability across governance regimes. In parallel, it can provide space for regional or thematic coalitions to develop and advance voluntary commitments, encouraging a ‘race to the top’ in good governance of AI.
The AI Dialogue can support capacity building, including for developing countries, by sharing good practices, technical expertise, policy tools and frameworks for public engagement. This ensures more equitable participation in global AI governance, and alignment of national actions with domestic democratic mandates, rather than concentrating influence in a few regions, or imposing one-size-fits-all models.
With an active Citizens Track, the AI Dialogue can also serve as a trust-building mechanism. By enabling open exchange on sensitive issues such as military AI use, data sovereignty, and algorithmic bias, it can help reduce geopolitical tensions, encourage responsible innovation, and develop public support for meaningful governance.
14. What are some of the existing initiatives, partnerships, or mechanisms that the AI Dialogue should build upon or connect with, and what added value could the AI Dialogue bring?
The proposal for a Citizens Track is about connecting the Global Dialogue with the many different public deliberations on AI already taking place across the world, and helping catalyse dialogues where they are missing. Active engagement from the Global Dialogue with citizen participation on AI will help bring coherence to distributed local conversations around AI governance: providing an incentive for groups to align on common questions, themes and actions. In turn, this supports education and capacity building with citizens on Global Dialogue themes, which can aid local mobilisation in support of actions arising from the Dialogue.
We also encourage engagement with existing efforts to support structured public engagement in AI, notably the Global Citizens Assembly and the Global Coalition for Inclusive AI. Both provide complementary approaches to support distributed public dialogue, and to aggregate insights to feed into future Dialogue processes.
The Global Citizens Assembly recently used sortition to convene 100 diverse participants for a COP30-linked climate assembly, producing expert-informed policy recommendations based on 40+ hours of participant learning. This model of structured transnational dialogue could effectively address key AI governance topics.
The global coalition on inclusive AI, launched at the 2024 Paris AI Action Summit, involves UN agencies, governments, civil society, and academia and is planning the roll out a global citizens dialogue on AI using deliberative polling in 100+ countries.
Clear commitments from the Global Dialogue to engage with distributed dialogue and with a Global Civic Assembly on AI would aid resource mobilisation and support stakeholder coordination.
We also highlight the importance of engaging with the MAP-AI network to maximise the inclusion of majority world civil society voices at the table during Dialogue meetings, and of engaging with the Open Government Partnership, as a mature mechanism for multistakeholder commitments and actions on governance.
15. How can different stakeholders contribute to the AI Dialogue? Please share recommendations for the format and structure of the AI Dialogue.
We note the draft agenda and structure of the Dialogue is now published, and so focus suggestions on measures that will work within the structure set out.
To support public voices at the first Global Dialogue, we suggest having routes to:
-
Showcase existing public voice. Providing exhibition space and side-event opportunities for evidence from existing public dialogues with communities affected by AI to be shown, including space for youth voice. Consider including space in the speaking list for plenary and breakout sessions for representatives of relevant prior citizen participation processes.
-
Connect with public dialogues. A number of national and local public dialogues on AI will be taking place across the time of the Global Dialogue. Consider facilitating side-event opportunities for live link-ups for citizens taking part in these to ask questions of experts and officials at the Dialogue would generate engagement and energy, and give Global Dialogue participants an opportunity to hear direct from citizens.
Over the long term, we recommend that future Dialogues explore docking points (pre-Dialogue input to agenda, slots on main agenda etc.) for a transnational Civic Assembly on AI Governance, to complement inputs from the Scientific Panel and other stakeholder groups.
A standing citizens track mechanism should help to close the feedback loop, tracking how public inputs have shaped Dialogue discussions, and providing accessible feedback to mobilised publics on Dialogue outcomes.
Funders play a critical role in shaping an inclusive Dialogue, and should engage in coordinated action to resource intermediary organisations capable of facilitating participation from underrepresented and non-digitally connected communities, ensuring that inclusion is not dependent on digital access alone but reflects real-world diversity of affected populations.
16. Which voices, communities, or perspectives are currently underrepresented in global discussions on AI governance? How could they be included?
The voices of diverse communities across the globe affected by AI: including communities who are not digitally connected or directly using AI, but whose education, employment, media, healthcare, environment and opportunities are already impacted by AI. Youth voices and voices of older people are specifically notable for their absence, but also are voices from indigenous communities, data labelling workers and women-led businesses.
Research shared at the Participatory AI Research & Practice Symposium (PAIRS) [1] alongside the India AI Impact Summit illustrated the challenges of relying only on digital forms of engagement and participation: addressing the need to work with community organisations that can facilitate ‘last-mile’ engagement, and outreach to communities affected by AI, but who would not be able to participate in online or highly formalised consultation processes.
Distributed community dialogue models are a powerful route to hear underrepresented voices and perspectives, as demonstrated in the recent ‘Generative AI in Education: Have Your Say’ distributed dialogue commissioned by the UK Department of Education in 2025. Through a toolkit for teachers to run classroom dialogues, including with pupils with additional educational needs, over 6 weeks the project surfaced critical views from more than 1000 students on the importance of maintaining student-teacher relationships as AI is introduced into schools, and led to changes in government standards for Generative AI in schools [2].
The UN Global Dialogue on AI Governance does not need to coordinate distributed dialogues, but by showing it will listen to them, as the UK Department for Education did for student input, it can catalyse much more inclusive and widespread engagement.
[2]: https://connectedbydata.org/projects/2025-dfe
17. What innovative engagement formats could most effectively foster meaningful and dynamic engagement during the AI Dialogue?
Thinking specifically about the 2026 Dialogue, and in addition to suggestions already made for a showcase of existing public voice on AI, we expand on the idea of live link-ups to citizen dialogues, and highlight the Hopes and Fears lab format.
The Dialogue could facilitate opportunities for live link-ups (Zoom etc.) with community assembly/participatory dialogue projects around the world - allowing citizens to put questions to a panel of experts participating in the dialogue.
This proposal draws on learning from the People’s Panel on AI - Deliberative Citizens Review organised alongside the UK AI Safety Summit in 2023 [1] where citizens who had learnt about the AI governance discussions taking place had the opportunity to question experts who had participated in the Summit in a side-event. This both helped citizens understand more the themes addressed by the summit, and to align their recommendations for action with summit debates - and gave an opportunity for the panelists to hear questions reflecting citizen concerns: helping ground their further engagement in the summit in consideration of citizen perspectives.
Another innovative format used on the fringes of the AI Safety Summit in 2023 was the ‘Hopes and Fears Lab’: a facilitated space designed for AI experts and members of the public to sit down together for short informal conversations on their respective hopes and fears for the future of AI. This can also act as a powerful way of grounding governance debates in everyday concerns and narratives: and breaking down barriers between stakeholders [2].
[1]: https://connectedbydata.org/projects/2023-peoples-panel-on-ai
[2]: https://www.kcesp.ac.uk/projects/the-hopes-and-fears-lab/
18. Please share examples of policies, practices, platforms, or approaches that promote effective AI governance or offer concrete solutions to addressing its challenges.
Our work offers concrete examples of effective participatory AI governance in practice.
The People’s Panel on AI (2023) demonstrated that members of the public with no particular prior experience of AI governance were able, over the course of four days, engaging with webcasts from the AI Safety Summit, fringe events, and experts involved in the Summit, to create a nuanced set of recommendations and messages that added depth to the discussions of the main Summit.
The Generative AI in Education: Have Your Say (2025) project demonstrates distributed dialogue influencing policy, providing framing questions and a toolkit of resources for self-organised community dialogues that both informed guidelines for safe AI in the classroom, and that supported governance action at the school level: aligning national and local governance conversations.
The Participatory AI Research and Practice Symposium (PAIRS) demonstrates the breadth of existing practice across the world on participatory governance of AI, from national-level consultations, cross-border discussions, and grassroots community conversations. It highlights the untapped potential of deliberative engagement on AI to build community agency and capability with respect to AI.
The Global Assembly on Climate and Food Systems, developed in collaboration with the Brazilian government hosting COP30, demonstrates how transnational deliberation can dock into global decision making, and how cross-language, cross-cultural deliberative engagement can make sense of complex topics and a range of community inputs.
The We The Internet (2021) project demonstrated how distributed dialogue can inform technology policy issues.
Building on all these projects, and working with a wide range of partners, we stand ready to support the development of a Citizens Track on AI Governance, showcasing existing citizen voice in 2026, connecting with and catalysing public dialogue as a continuity mechanism towards the 2027 Dialogue, and developing mandates and frameworks for embedded transnational citizen deliberation in future Dialogue iterations.
