Governing the Algorithmic State: A Systematic Literature Review of Artificial Intelligence in Public Administration (2018–2026)

The integration of artificial intelligence (AI) into public governance has transitioned from speculative administrative foresight to an active operational imperative. This study presents a comprehensive Systematic Literature Review (SLR) synthesizing empirical, theoretical, and institutional evidence on AI applications across public administration over a nine-year temporal window (2018–2026). Following the PRISMA 2020 protocol, an initial corpus of 7,865 records retrieved via an audited computational pipeline from the OpenAlex database was screened, yielding 60 high-impact studies strictly published in SCImago Q1 journals. Bibliometric and thematic analyses reveal an exponential acceleration in scholarly output post-2022 (86.7%), yet demonstrate persistent domain skew: 56.7% of studies treat AI as a generalized, technology-agnostic construct, and generative AI remains underrepresented (1.7%). Thematic synthesis maps the transformation of frontline discretion from Lipsky’s street-level bureaucracy to screen-level and system-level bureaucratic paradigms. Anchored in public management theory, we propose an Integrative Socio-Technical Governance Framework and articulate three theoretical propositions (P1–P3) bridging administrative discretion, algorithmic accountability architectures, and administrative burden mitigation.