Algorithmic Accountability and the Rise of AI-Driven Governance in Public Administration

Abstract

The diffusion of artificial intelligence (AI) technologies into public administration has fundamentally altered how governments make, justify, and review decisions that affect citizens' lives. From welfare eligibility screening to tax scrutiny selection, from judicial case management to law-enforcement risk assessment, algorithmic systems now mediate a growing share of state power. This shift raises urgent questions about algorithmic accountability — the capacity of citizens, courts, legislatures, and oversight bodies to understand, contest, and correct decisions made or substantially shaped by automated systems. This paper undertakes an analytical review of the theoretical foundations, global regulatory trajectories, and empirical experience of AI-driven governance, with particular attention to the Indian administrative context. Using a qualitative, document-based analytical methodology, the paper examines landmark international scholarship on algorithmic accountability alongside Indian case studies including the Aadhaar biometric identification system, the Faceless Assessment Scheme of the Income Tax Department, the Supreme Court's SUPACE and SUVAS initiatives, and NITI Aayog's Responsible AI policy architecture. The analysis finds that while algorithmic governance offers demonstrable gains in efficiency, consistency, and corruption reduction, it simultaneously produces new forms of opacity, exclusion, and diffused responsibility that existing administrative law doctrines are ill-equipped to address. The paper argues for a hybrid accountability model that combines technical auditability, human-in-the-loop safeguards, participatory design, and statutory rights to explanation, and it situates India's emerging regulatory architecture — particularly the Digital Personal Data Protection Act, 2023 and NITI Aayog's Responsible AI framework — within this global conversation. The paper concludes with policy recommendations for strengthening accountability without foreclosing the developmental promise of AI in governance.