Administrative discretion and algorithmic decision-making: governing public authority in the algorithmic state

Abstract

Algorithmic decision-making is increasingly used by public authorities to structure administrative processes and to enhance efficiency and consistency. Existing debates largely evaluate automation in terms of accuracy, transparency, and compliance, while paying limited attention to its institutional implications for administrative governance. This article examines how algorithmic systems reshape administrative discretion within public decision-making. It argues that discretion is not eliminated by automation but redistributed across socio-technical infrastructures, including system design, data practices, and organisational implementation. This redistribution generates governance challenges for public authorities, particularly concerning attribution of authority, accountability gaps, and the limits of administrative and judicial review. By conceptualising administrative discretion as an institutional safeguard, the article explains why effective algorithmic governance requires mechanisms that preserve responsibility, justification, and oversight within digital administration. It concludes by outlining regulatory and institutional reforms through which administrative law can govern decision infrastructures while maintaining its core rule-of-law commitments.