Model risks in the financial sphere under the conditions of the use of artificial intelligence and machine learning

Objective: within the framework of RegTech and SupTech technologies, to assess the transformation of model risks and ways to minimize them under the increasing use of artificial intelligence methods.Methods: a systematic approach to the analysis of the quality of economic models. Historical, logical, and statistical methods of research.Results: the Russian and foreign experience of accounting for model risks in the financial industry is considered. Theoretical and practical works on the regulation and management of model risks in the activities of financial sector organizations are studied. The role of machine learning and artificial intelligence technologies in solving the modern problems in the functioning and regulation of financial organizations is determined. The key model risks are considered, as well as the directions of changing their specifics as a result of the artificial intelligence technologies development, primarily machine learning, and increasing the capabilities for storage and transmission of a large amount of data. The main methods of data processing and model construction are considered, as well as their advantages in terms of reducing model risks. It is determined that the reduction of model risks using RegTech and SupTech technologies is possible due to the development of artificial intelligence technologies, which will require, among other things, the elaboration of the appropriate legal field.Scientific novelty: the unique feature of the article is a comprehensive consideration of the problem of model risks in the finance industry and of the impact of artificial intelligence technologies on them in mathematical, legal, economic aspects, as well as the description of the situation in this area both abroad and in Russia.Practical significance: the information presented in the article can be used by regulatory authorities and commercial banks in the tasks related to minimizing specific model risks in their activities.