Tim presented to country offices from the United Nations Development Programmes Europe and Central Asia (RBEC) region, as part of a series on ‘AI in Practice’.
This hour long session explored why public participation matters in the design, deployment, and governance of AI systems. It looked at practical approaches governments can use to involve citizens and stakeholders, strengthen accountability, and build public trust around AI in the public sector.
The session write-up from UNDP reflected on the following key insights from the presentation and discussion:
1) Public participation is essential to build trust, better AI systems, public legitimacy, and people’s agency: The session highlighted four core reasons for engaging people in AI decisions: building trust in AI, improving system quality, securing a public mandate for AI-driven reforms, and strengthening people’s ability an agency to shape how AI affects their lives. Public participation is not just consultation - it helps surface concerns, improve governance, and ensure AI reflects real societal needs and values.
2) Meaningful participation must influence real AI decisions, not just gather feedback: Tim emphasized that good AI governance should be collective and deliberative, where people have time, information, and real influence over outcomes. Many current approaches rely on surveys or individual consent, while key decisions about AI purpose, policy and deployment remain closed. Effective participation requires people to help shape what AI is used for and how it is governed.
3) Public engagement can shape AI across its lifecycle, from design to policy to accountability: The session showed that participation is relevant at multiple decision points, including shaping AI model behaviour, guiding AI use in public services, informing national AI policy choices, and enabling communities to identify and respond to AI harms and bias. This highlights participation as an ongoing process, not a one-off consultation.
4) Practical participation methods exist, but institutional foundations are needed to sustain them: Concrete approaches such as citizen assemblies, distributed dialogues, collective intelligence exercises, and community-led audits were shared. The session also stressed that meaningful participation requires leadership support, clear standards, institutional structures, and partnerships with affected communities to be effective and scalable in government contexts.
The presentation also includes a range of concrete examples, methods, and links to resources that colleagues can explore further when thinking about how to design participatory AI processes in their contexts. One tool highlighted in the session was the Good Governance Game (https://connectedbydata.org/game), which offers method cards and exercises to help teams design participatory governance processes around AI and data decisions.
If you would like an introduction to public participation in AI for your organisation or network, mailto:tim@connectedbydata.org.