Algorithmic Governance in Public Administration: An Analysis of the Challenges of Accountability, Transparency, and Responsibility in Artificial Intelligence–Based Decision-Making
Type: article
Abstract: The increasing use of artificial intelligence in public administration has transformed the structure of governmental decision-making and created new legal challenges concerning transparency, accountability, and responsibility. Algorithmic systems are increasingly employed to classify individuals, assess risks, allocate public resources, support regulatory enforcement, and assist or automate administrative decisions. Although these technologies may improve efficiency, consistency, and institutional capacity, they may also weaken traditional safeguards of administrative law when their operation is opaque, difficult to explain, or distributed among multiple public and private actors. This article examines the legal implications of algorithmic governance in public administration with particular emphasis on the principles of legality, procedural fairness, transparency, explainability, equality, human oversight, judicial review, and institutional responsibility. Using a descriptive-analytical legal approach, the study evaluates the extent to which existing public-law principles can address the challenges created by AI-based decision-making and identifies areas in which additional regulatory mechanisms are required. The analysis shows that transparency must extend beyond simple disclosure of algorithmic use and should include meaningful explanations of the factors influencing individual administrative decisions. It also argues that public authorities must remain legally responsible for decisions generated or materially influenced by artificial intelligence, even where systems are developed or supplied by private contractors. Effective human oversight, algorithmic impact assessments, continuous auditing, accessible complaint procedures, and meaningful judicial review are essential for preventing accountability gaps. The article concludes that a sustainable model of algorithmic governance requires the integration of technological innovation with established principles of public law. Artificial intelligence should support, rather than displace, legal responsibility, and the use of algorithmic systems must remain subject to identifiable lines of explanation, review, and institutional accountability.
Author: Karami, Mohammad Reza
Institution: Shahid Beheshti University
Language: English
Subfield: Law
Field: Social Sciences
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W7223195553