Correlation and differentiation of public law institutions of artificial intelligence and automated administrative procedures

Abstract

Purpose. The purpose of this study is to determine the place of the emerging institute of artificial intelligence (AI) in the system of public law institutes and to identify its correlation with the institute of administrative procedures in relation to the preparation and adoption of individual (non-normative) administrative acts by executive authorities and other public administration bodies. The subject matter is deliberately limited to positive administrative proceedings, including licensing, registration, public service and certain non-delict control procedures. Proceedings on administrative offences and administrative litigation are used only as adjacent material for analysing the reasoning and reviewability of an act and are not equated with administrative procedures.Methods. The research is based on general scientific methods (analysis, synthesis, systematic approach) and specialized legal methods – formal-legal, comparative-legal, systematic-structural. The formal-legal method was used to study the regulatory legal acts of the Russian Federation governing the creation and application of AI in the public law sphere. The comparative-legal method was used to compare the Russian approach with the approach of the European Union as an international association that adopted the first systematic act on AI regulation. The systematic-structural method made it possible to determine the place of the AI institute in the system of public law institutes.Results. It was established that the AI institute in public law is complex not because AI rules are mechanically dispersed among different legal acts, but because the regulated relations, special concepts, common guarantee function and need for cross-sectoral coordination are unified. The correlation between the AI institute and administrative procedures is analysed in relation to individual administrative acts. Two types of legal defects are identified. The first is a legal gap: the absence of special rules governing the preparation of such an act with the use of AI and the review of the unreasonableness of an algorithmic conclusion. The second is a legal conflict in a broad sense: the tension between the requirement to give reasons for an act and the technical opacity of some machine-learning models. A distinction is made between automated recording of legally significant events (Article 28.6 of the Code of Administrative Offences of the Russian Federation) and automated preparation of an administrative decision.Conclusions. The AI institute in public law is an emerging complex cross-sectoral institute. Its correlation with administrative procedures requires the consolidation of two groups of rules: procedural rules on indicating the fact of using AI in the preparation of an individual administrative act, on requirements for algorithmic transparency and on the possibility of challenging an algorithmic conclusion; and classification rules distinguishing permissible and impermissible areas of AI application in public authority activities according to the level of risk. The conclusions directly concern individual administrative acts; the use of AI in departmental rule-making requires separate research.