Artificial Intelligence and the Future of Citizen Participation

5. Conclusions

Governments are using AI tools to improve citizen participation processes

Across OECD member and partner countries, evidence shows that citizen participation is emerging as a prominent area of experimentation in the use of AI in government. The analysis of 50 case studies described in this report confirm that AI tools open significant opportunities to improve and enhance citizen participation processes. For instance, AI-powered sense-making, information development, translation, transcription, virtual assistance, facilitation, simulation, and architecture tools can create new bridges, and empower citizens in navigating complexity in policy making, while expanding government capacity to take citizen voices into account when making decisions. Various applications can concur in supporting the activities of practitioners and policymakers to address some of the challenges of participation, while supporting citizens in playing their part All in all, AI tools can improve the design and the accessibility of citizen participation processes, enable participation at scale while maintaining the quality, and allow for a better allocation of government resources. The Typology proposed in this report is meant to support policymakers and practitioners in navigating the opportunities offered by AI in this field.

Among the different applications analysed, AI-powered sense-making, virtual assistance and translation tools are currently the most widespread, often integrated in existing digital tools and platforms for participation. Some citizen participation processes, such as public consultations and representative deliberative processes, are areas of particular interest with respect to the opportunities offered by AI tools. The adoption of AI tools for participation is for the time being mostly driven by local-level initiatives, showcasing the role of local governments in pushing the boundaries of democratic innovations. However, some countries are leveraging their national innovation labs and international ecosystems to develop and adopt AI tools to support participation at the national level.

Beyond task-specific tools, technological progress is creating room for further experimentation in AI-human policymaking. Beyond citizen participation processes, evidence shows that AI tools can support all areas of open government, including transparency and access to information, the protection of civic space, government accountability, and government communication.

Governments can act to steer the trustworthy adoption of AI while continuing to improve citizen participation

Most initiatives analysed adopted process-specific tools and implemented processes on an ad hoc basis. While these cases highlight the potential for AI to enhance citizen participation, they also underscore the need for governments to engage with this trend strategically.

AI systems entail significant risks on multiple levels: (1) ethical risks resulting in adverse outcomes, threats to the civic space and surveillance, lack of transparency and explainability, and affected quality and depth of dialogue and deliberation; (2) operational risks entailing overreliance on AI systems and privacy concerns, (3) exclusion risks that might create new barriers to participation for some segments of the population, (4) public resistance risks related to citizens’ mistrust in the use of AI for participation, (5) risk of inaction and missed opportunity of not adopting AI when relevant. Governments are also confronted with implementation challenges when adopting AI tools for participation: (1) limited skills among civil servants, (2) limited actionable guidance, (3) difficulties in scaling pilot experiences, and (4) market concentrations of AI tools and lock-ins once tools have been procured. Governments should act to address these risks and overcome these challenges by creating the conditions to design and select fair and transparent AI tools and build capacity and steer ecosystems to seize the opportunities offered by this technology, while preserving the quality of citizen participation processes and protecting civic space.

Furthermore, AI and technology alone cannot solve all challenges of citizen participation. Governments should continue working to improve the quality and the accessibility of citizen participation processes while establishing mechanisms to ensure that citizens’ efforts have a clear impact on policy and decision-making. AI and other digital and emerging technologies should be seen as supporting tools to further improve and enhance meaningful and impactful citizen participation.

Finally, the intersection between citizen participation and AI should be seen as a two-way street: on one hand, AI tools can support and enhance citizen participation processes. On the other hand, governments can leverage citizen participation practices to involve citizens in the design, deployment and monitoring of AI systems.

Looking ahead, governments should move beyond short-term experiments and task-specific applications to consider more fundamentally how democratic institutions and participation processes should evolve in an AI-enhanced context, particularly as technologies rapidly progress towards more advanced general-purpose capabilities.