DOI: 10.67733/rlipe.2.1.4 ISSN: 3062-4487

The Problem of Fault Attribution in AI-Supported Administrative Acts and Activities

Zeynep Çelik Gülseven
Show PDF Cite
The increasing use of artificial intelligence technologies in public administration necessitates a reassessment of administrative liability. Traditional theories of administrative liability are based on the assumption that administrative acts are adopted by a specific public official or administrative authority and that, therefore, a decision can be attributed to a specific individual or institution. However, in AI-supported decision-making processes, numerous actors, such as data providers, software developers, algorithm designers, system operators, and public authorities, become part of the decision-making mechanism. In addition, the ability of machine learning-based systems to develop autonomous behaviors reduces the transparency of the decision-making process and complicates the identification of the source of harm. This study examines the problem of fault attribution arising from AI-supported administrative acts and activities. The main argument of the study is that artificial intelligence is not a factor that eliminates liability; rather, it is a technological phenomenon that complicates the determination of fault, the establishment of causation, and the attribution of liability to specific individuals or institutions. Within this framework, the study first examines how artificial intelligence transforms administrative acts and activities. It then analyzes the problem of fault attribution and the role of the administration in this regard. Finally, the effects of AI systems on administrative liability are evaluated from the perspective of fault.

More from our Archive