1. ALGORITHMIC REASON-GIVING, ARBITRARY AND CAPRICIOUS REVIEW, AND THE NEED FOR A CLEAR NORMATIVE BASELINE.
- Author
-
Averill, Cameron
- Subjects
MACHINE learning ,PATENT offices ,SOCIAL Security (United States) ,FOOD stamps ,ROBOTIC process automation ,NOTICE of proposed rulemaking (Administrative law) ,DRUG approval laws - Abstract
The text explores the use of algorithms by federal agencies in administrative processes, emphasizing the importance of policies and procedures in guiding decision-making. It discusses the potential for bias in datasets used to train algorithms and the need for transparency in bias mitigation strategies to build trust in algorithmic results. The author proposes a normative framework to evaluate algorithms based on systemic reasons, intended recipients, decision nature, and bureaucratic considerations, suggesting that algorithms used for agency action should disclose systemic reasons to survive review. The article aims to reconcile the use of algorithms with administrative law's reason-giving requirements and highlights the challenges and implications of using algorithms as reasons for agency action. [Extracted from the article]
- Published
- 2024