Algorithm in the Structure of Administrative Decision-Making: A Model of Explainability and Burden of Proof for an Effective Legal Remedy within Discretionary Power in Public Administration

The introduction of artificial intelligence (AI) systems into public administration decision-making processes poses fundamental challenges to procedural guarantees, particularly the right to an effective legal remedy. This article examines how the use of algorithms, especially those of a “black box” nature, affects the transparency of proceedings and the possibility of judicial review. Based on a dogmatic analysis, rooted in the concept of a rationalized administrative decision, and a comparative case study analysis (Poland, the Netherlands, Estonia, Finland), the author argues that the lack of explainability in AI systems paralyzes the right to appeal. In response to the diagnosed problems, the article proposes a model for a minimum standard of “algorithmic justification”, which can be implemented within the legal system. This model aims to restore transparency, reverse the unfavorable burden of proof for the individual, and adapt the judicial cognition model to the new technological reality, taking into account ESG frameworks as a standard of due diligence for public authorities.