Collaborative Research: HCC: Designing Civic AI to Connect Residents, Communities, and Local Governments

Cities collect large amounts of information about public services, including 311 reports, open datasets, and records from public meetings. Currently, this information is often hard for residents to find, understand, and use effectively. Citizens report problems, such as street flooding, parking, unsafe streets, or limited mobility options, without knowing whether others have raised the same issue, without feedback on the city’s actions, and without information on how they might take part in decisions or actions around these reports. This project studies how artificial intelligence (AI) can help turn scattered civic information into actionable community knowledge. The goal is to help residents, community organizations, and local governments better understand public problems, connect individual reports to broader patterns, and identify practical next steps. The work will support more transparent, responsive, and accountable public services, while also training students and strengthening partnerships among universities, local governments, and community groups in New Orleans and Pittsburgh. This project will develop and evaluate community-centered civic AI systems that support public participation, shared sensemaking, and city communication around civic data. The team will conduct participatory design activities with residents, civic organizations, people with disabilities, and municipal staff to identify what successful civic AI should do, what kinds of information it should make visible, and what privacy and accountability concerns it must address. Building on these activities, the project will develop and study AI tools that connect resident questions and complaints to relevant evidence, including prior service reports, public meeting records, policy discussions, community contributions, and opportunities for participation. These tools will use source documentation, grounded summaries, uncertainty information, and methods for preserving different viewpoints so that the system does not present a single consensus answer that might leave out important factors. The project will also deploy an interactive civic dashboard and social platform that allows residents, organizations, and city staff to view, discuss, contest, and update shared civic knowledge over time. The research will evaluate both technical performance and social outcomes, including whether the systems help people understand civic issues, find relevant meetings or organizations, contribute local knowledge, calibrate trust in AI outputs, and identify meaningful next steps for action. The project will produce design principles, evaluation methods, reusable tools, and evidence about how AI can support civic participation without flattening community disagreement or replacing public deliberation. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.