Artificial intelligence and discretionary decision-making in construction offense cases - an empirical study

Abstract

The paper deals with the issue of integrating artificial intelligence into discretionary decision-making. It focuses on decision-making within the framework of sanctioning construction offences, since in this area a strong trend of electronic decision-making processes has recently been evident within the framework of recodification processes. However, these decision-making processes also represent a special challenge in terms of input information and in terms of existing data sets on application practice in sanctioning. Based on the scientific literature and the projection of the knowledge thus obtained into the Slovak reality, it discusses the benefits and strengths and the negatives and risks of decision-making or participation in decision-making by artificial intelligence. In the next part, the paper presents the results of an experiment aimed at testing how artificial intelligence models commonly available to the public can deal with discretionary decision-making in sanctioning cases. It tests how - based on a description of the facts of a fictitious case supplemented with basic information about the relevant legal regulation - artificial intelligence proposes a decision on the imposition of a sanction and how consistent it is in doing so. Based on theoretical and empirical foundations, the paper summarizes the benefits and risks of discretionary decision-making in sanction matters by artificial intelligence.