Abstract

This study aimed to examine the nature of administrative decisions based on artificial intelligence algorithms, algorithmic bias in administrative decisions, and the legal safeguards established to mitigate such bias. The study adopted the analytical approach, supplemented by the comparative approach to the extent required by the nature of the research. The study examined the nature of administrative decisions based on artificial intelligence algorithms by analysing their concept, constituent elements, and the legal basis for their adoption by administrative authorities. It further examined algorithmic bias in terms of its concept, characteristics, and sources before analysing the legal safeguards established to mitigate such bias. The study reached several findings. Most notably, it found that the use of artificial intelligence algorithms in issuing administrative decisions does not alter their legal nature; rather, such decisions remain administrative decisions governed by the general legal rules applicable to administrative decisions. It further found that algorithmic bias constitutes one of the principal legal challenges associated with administrative decisions based on artificial intelligence algorithms and may arise from multiple sources, including data, system design, human intervention, and system operation. The study also concluded that the Saudi legal framework provides a general legal basis and safeguards that may be relied upon in the use of artificial intelligence algorithms. However, it does not establish a dedicated legal framework governing administrative decisions based on artificial intelligence algorithms, whereas the European Union's General Data Protection Regulation (GDPR) provides specific safeguards for decisions based on automated processing, and French legislation provides a more detailed regulatory framework for administrative decisions based on algorithms. The study concluded with several recommendations, most notably the enactment of a dedicated legal framework governing administrative decisions based on artificial intelligence algorithms that reflects the specific nature of such decisions; expressly requiring administrative authorities to disclose the use of algorithmic processing in administrative decision-making; granting the person concerned the right to request that the administrative decision be reviewed by a competent official, to present his or her views, and to object to the decision where it affects his or her rights or legal position; and prohibiting the issuance of an individual administrative decision based exclusively on an algorithm capable of autonomously modifying its operating rules or decision-making criteria, thereby ensuring effective human oversight and enabling the review and verification of the lawfulness of the decision.

Keywords

Administrative Decision, Artificial Intelligence Algorithms, Administrative Decisions Based on Artificial Intelligence Algorithms, Algorithmic Bias.