Artificial Intelligence in Public Administration: A Bibliometric Analysis from 2015-2026 Based on Scopus Data

Background: The rapid advancement of artificial intelligence (AI) has driven a significant transformation in modern public administration practices; however, comprehensive bibliometric mapping focused on the explosive period of the last decade remains highly limited. Objectives: This study aims to analyze the developmental landscape of scientific literature on AI in public administration during the 2015-2025 period, encompassing annual publication trends, geographical and institutional distributions, international collaboration networks, as well as the thematic and conceptual structures dominating academic discourse in this field. Methods: This study employs a quantitative bibliometric approach combined with a systematic literature review based on the PRISMA framework. Data were retrieved from the Scopus database using advanced search queries with the keywords "artificial intelligence" AND "public administration," covering publications from 2015 to 2025. Following a rigorous selection process based on inclusion and exclusion criteria, 729 English language journal articles were analyzed using VOSviewer software to generate visualizations of keyword co-occurrence networks, international collaborations, and thematic cluster mapping. Results: The findings indicate significant growth in publications post 2020, with the United States emerging as the dominant contributor. Co-occurrence analysis identified three primary thematic clusters: digital transformation and governance, data-driven approaches for decision-making, and the application of AI in public health, while issues regarding algorithmic ethics and public trust remain underrepresented. Conclusion: AI in public administration is a rapidly evolving field of study that requires a more inclusive, contextual, and research-oriented agenda focused on democratic values and responsible governance.