Administrative Liability for Damages Caused by Artificial Intelligence Systems in Public Services: An Analytical Study in Light of the Principles of Legality and Transparency

This research examines the issue of administrative liability for damages caused by artificial intelligence (AI) systems in public services. It offers a legal analysis of legality and transparency in automated administrative decisions. The topic gains importance as public administrations increasingly rely on intelligent systems, raising unprecedented legal challenges particularly with respect to oversight and accountability. The study is structured into three main sections: the first outlines the theoretical framework of administrative liability and AI; the second explores the compatibility of automated decisions with the principle of legality and the scope of judicial oversight; and the third examines the components of administrative liability, supported by comparative judicial models from France, the United States, and Egypt, with an analysis of their applicability within the Omani legal context. The study concludes that administrative liability may arise from supervisory negligence or algorithmic mismanagement-even in the absence of direct technical error. It recommends the enactment of legislation to regulate the use of AI in public administration, promote algorithmic transparency, and institutionalize the principle of “meaningful human oversight” as a legal safeguard for fair accountability.