Artificial Intelligence and the Future of Citizen Participation

Executive summary

This report explores the opportunities and challenges of the adoption of artificial intelligence (AI) tools to improve citizen participation. To do so, it builds on desk research and analysis of 50 AI use cases in citizen participation processes from 22 OECD member and partner countries. It proposes a Typology of applications which aims at supporting government officials and practitioners navigate the landscape of AI tools for participation based on their needs and the challenges they face. The report provides insights on emerging trends in the adoption of AI tools for participation, analyses the related risks, and outlines relevant mitigation strategies that allow governments to steer the trustworthy adoption of the technology.

Key findings

AI offers a wide range of opportunities in supporting citizen participation

AI tools can support governments and practitioners addressing some of the challenges they face in conducting citizen participation efforts by (1) improving the design and accessibility of citizen participation processes, (2) enabling the implementation of processes, including interpretation of results, at scale while preserving their quality, (3) communicating better on citizen participation processes, bridging the gap between participants and the broader public debate, and (4) reducing costs associated with citizen participation processes. This report proposes a Typology to navigate these opportunities. The Typology identifies nine types of applications of AI tools to support and improve citizen participation processes: Information development, Sense-making, Translation, Transcription, Virtual Assistance, Moderation, Facilitation, Simulation and Architecture. These applications support government officials, practitioners, and citizens in the various tasks and activities required to implement and take part in participatory and deliberative processes. In some cases, AI tools can enable new formats of participation altogether, or at scales not feasible without AI. For each application, the Typology explores the relevant uses of AI tools in both the front-office (the citizens-government interface) and the back-office of government (governments’ internal operations). Finally, each application is analysed against its relevance in addressing the existing gaps and challenges to achieving meaningful and effective participation.

AI-powered sense-making, virtual assistance (e.g. personalised support to participants through conversational interfaces), and translation tools stand out as the most widespread. Among the different types of citizen participation processes, deliberative processes constitute a lively field of experimentation with AI tools, in particular to improve live facilitation activities and to enable deliberation at scale.

Beyond citizen participation processes, AI presents relevant opportunities to support the implementation of broader open government policies and practices, including access to information, government communication, and the protection of civic space. New and future developments in the technology encourage a reflection on more cutting-edge and emerging practices at the intersection of AI and participation, such as possible uses of AI agents in citizen participation processes and human-AI policymaking.

Governments around the world are already using AI tools to design and implement citizen participation processes taking place at all levels of governance. However, like other types of government AI efforts, the adoption of AI mostly occurs on an ad-hoc basis or through pilot initiatives. While the local government level drives this emerging trend, a growing number of national governments are developing coordinated strategies or tools to integrate AI in their participatory processes.

The use of AI in citizen participation entails risks and implementation challenges

Deploying AI tools in citizen participation processes does not come without risks and challenges. Building on the analysis carried in the OECD Report Governing with Artificial Intelligence: The State of Play and Way Forward in Core Government Functions, this report distinguishes between risks and implementation challenges and focuses on their respective implications in the use of AI applied to citizen participation. With respect to risks, it looks into:

• Ethical risks, such as inadequate or skewed data in AI systems, leading to inaccurate of adverse outcomes, the lack of transparency and explainability of some AI tools, the misuses of AI that might undermine the civic space, and the effects of all previous elements on the depth and quality of participation and deliberation.

• Operational risks, such as overreliance on AI systems, hallucinations, privacy and data concerns, and cyber threats, which might result in reduced quality of participation and further erosion of public trust.

• Exclusion risks, including the danger of widening digital, language, and societal divides.

• Public resistance risks, including the potential impact on citizens’ trust.

• Inaction risks, corresponding to the missed opportunities of non-adopting AI system when relevant.

Beyond risks, governments also face organisational and institutional challenges. These are related to the integration of AI in government operations. These challenges include existing gaps in skills and literacy, overreliance on AI systems, inadequate investments, and scaling of practices and tools. Key recommendations

This report suggests a list of actions for governments to create the right conditions to benefit from the opportunities offered by AI tools to improve citizen participation processes while mitigating the risks. Consistent with the OECD Framework for Trustworthy AI in Government, governments should:

• Establish guardrails to anticipate and manage risks associated with AI adoption in citizen participation, including: adopting fair and transparent AI systems and associated governance and processes by ensuring compliance with democratic values, the rule of law, and the protection of the civic space, closing data divides, and consider adopting open-source, open standards and open weights models when relevant. Moreover, governments should act to prevent exclusion by offering low-tech alternatives for citizens with limited digital access while improving AI literacy. Finally, governments should be transparent in their use of AI in participation to prevent public resistance.

• Strengthen relevant enablers to overcome key implementation challenges of AI systems for citizen participation, including by raising awareness and building skills within government on how to use AI responsibly when designing, implementing, and participating in citizen participation processes.

• Foster engagement with citizens stakeholders in shaping and governing AI systems for participation, including by leveraging citizen participation processes to involve citizens in AI development and governance, strengthening innovation ecosystems to support co-creation platforms and communities of practice, and scaling successful pilots.

• Continue to work towards more meaningful and impactful citizen participation by addressing its institutional and organisational challenges and by improving the design and accessibility of processes beyond the use of digital tools, AI, and other emerging technologies.