AI for democratic participation? Citizens’ perceptions of problems in online participation processes and AI-based solutions

This study investigates citizens’ perceptions of problems in online citizen participation (OCP) and their views on the potential of artificial intelligence (AI) to address them. Based on a nationwide survey in Germany (n = 484), we identified six categories of perceived problems, including lack of binding consequences, issues related to the usability and functionality of OCP platforms, and deficits in tolerance and civility. Participants considered three clusters of AI applications particularly relevant: detecting norm violations, structuring discussions, and fostering activity. While citizens valued AI for filtering hate speech, bots, and misinformation, they were skeptical toward interventions that directly steer debates. Regression analyses show that prior participation experience and external political efficacy are linked to lower problem perceptions. Strikingly, higher problem perception correlates with less support for AI solutions, suggesting skepticism about technology’s ability to remedy democratic shortcomings. The findings highlight opportunities and limits of AI in strengthening OCP.