When Algorithms Refuse: The Right To A Reasoned Government Decision In The Age Of Artificial Intelligence
Sideque N, B.S. Abdur Rahman Crescent School of Law
ABSTRACT
Governments increasingly use artificial intelligence to support decisions on welfare, taxation, licensing, immigration and public security. These systems may improve administrative efficiency, but they can also make it difficult for an affected person to understand why a decision was taken. This paper examines whether an AI-assisted government decision can satisfy the legal duty to give reasons. It analyses the black-box problem, the implications of opaque automated decision-making for equality, privacy, fair procedure and effective remedies, and the allocation of responsibility between public authorities and private technology providers. Drawing on international human-rights instruments and comparative case law from the Netherlands, the United Kingdom, the United States and the European Union, it argues that a legally adequate reason must be case-specific, intelligible and capable of challenge. The paper proposes mandatory human review for high-stakes decisions, case-specific explanations, independent audits, discrimination testing, proper records, effective appeals and clear institutional accountability.
Keywords: artificial intelligence; automated decision-making; right to reasons; algorithmic transparency; human rights; administrative law.
