The Embedding of Fairness Value in Algorithmic Administrative Discretion

Xin Wang

Keywords: Algorithmic Administrative Discretion, Fairness, Rule of Law, Algorithmic Governance, Administrative Law

Paper Type: Full papers

Abstract: This paper examines how fairness, as a foundational value of administrative law, can be embedded within algorithmic administrative discretion. Algorithmic administrative discretion systems increasingly play an assistive role in public administration and, in certain domains, fully automate human decision-making processes,thereby transforming discretionary judgment into processes governed by data, models, and human–institutional rules. Drawing on administrative law theory and computational governance theory, the research develops a conceptual framework for algorithmic administrative discretion, identifying its structural layers—data, model, and human–institutional—and its dynamic operation through information, decision, and feedback flows. Fairness is analyzed as a normative logic internal to algorithmic discretion rather than as an external compliance requirement. Its institutional realization is articulated through lifecycle governance encompassing three stages: ex ante design and legality review, in-process human–algorithmic oversight, and ex post accountability and redress. Embedding fairness in this manner ensures that digitalized administration remains aligned with legality, proportionality, and procedural justice, thereby safeguarding the rule of law in the age of automation.

Submission Number: 7