Citizen-Centric Approaches to AI in Government Programs: Lessons From Experimental Studies

Abstract

Emerging technologies have far-reaching effects on public governance. Artificial Intelligence (AI) is one such technology. Yet, given concerns of privacy, accountability, and bias with AI use, public managers grapple with deciding if, and to what extent, AI should be included in delivering public services. Current research has made attempts to unpack the determinants of AI use in these settings, but with greater focus on the technology rather than its users. The current study shifts focus to citizens' perceptions and asks what the conditions are under which citizens would embrace high-performing AI government service delivery. The study leverages a sample of 612 U.S. residents in two experiments that investigate public sentiments towards AI in public service contexts where it is implemented well. Findings reveal citizens' nuanced support for AI in simplifying complex programs delivering tailored services and reducing fraud. These insights can guide decision-making on AI use in public service delivery.