Annotation

The paper examines the legal limits of the use of artificial intelligence technologies in courts on the example of administrative proceedings. Based on foreign regulations, current legislation, and judicial practice in Russia, it considers the risks arising from the algorithmization of public law disputes: the “black box” problem in relation to judicial evidence, the risks of algorithmic discrimination, and the threat of erosion of the principle of the court's active role are analysed in detail. The author's model of procedural guarantees is proposed: classification of cases according to the admissibility of automation, presumption of distrust towards the algorithm, reverse burden of proof when challenging decisions involving AI elements, the party's right to a meaningful explanation of the algorithmic decision, and the institution of independent algorithmic audit.