What is civic participation in artificial intelligence?

There are increasing calls across disciplines and sectors that the public should participate in decisions about the use of artificial intelligence (AI). Public input in governmental decision-making is particularly crucial to promoting a well-functioning democracy and mitigating harms from AI. However, AI’s opacity, mutability, and resource requirements impede meaningful civic engagement particularly in urban environments. Many prior systematic reviews of civic participation and AI draw on the smart city literature. However, several other disciplines influence civic participation in AI so a siloed disciplinary focus offers only partial guidance for participation’s future role in AI. Our multi-disciplinary analysis blends works in smart cities, and in public policy, communication and, importantly, computer science to reveal distinct and highly variable pathways for civic participation. We use a sequence of manual and automated steps to conduct a structured literature analysis beginning with over 3,000 articles. We categorize authors’ work on participation in AI into five themes: participation as a natural byproduct of automating government, participation facilitated through the medium of AI, participation in AI as quantification, participation as a technocracy of trust, and participation as meaningful. With few exceptions, authors seemed not to challenge the status quo nor diminish the authority of the experts. Authors focused on the processual without the influence and AI aided in that process orientation. We conclude that the future of public participation in AI requires careful attention to become meaningful including recognition of neoliberal intent and power differentials.