Artificial Intelligence Adoption in Public Administration: Challenges for Accountability, Transparency, and Public Service Delivery

Abstract

This study examines artificial intelligence adoption in public administration by focusing on its challenges for accountability, transparency, and public service delivery. The objective is to analyze how AI-based systems reshape administrative responsibility, decision visibility, frontline discretion, and citizen-oriented services. The study employs a qualitative method with a case study design because AI adoption is a context-dependent governance phenomenon that requires in-depth interpretation of institutional practices, actor perceptions, and implementation dynamics. The research was conducted at a metropolitan local government digital transformation agency in Indonesia, selected because it has introduced AI-supported complaint management, service monitoring, and administrative decision-support initiatives. Data were collected from twelve purposively selected informants, including senior officials, technical officers, legal and audit staff, frontline service officers, a civil society representative, an academic expert, and a public service user. The selection was based on their direct knowledge of AI implementation and public service governance. The findings show that AI improves service speed, workload management, and data-based monitoring, but also creates accountability ambiguity, transparency deficits, uneven staff readiness, and risks to human discretion. The study recommends clear accountability mapping, explainable AI procedures, public disclosure, human oversight, staff training, and citizen appeal mechanisms.