METHODS OF DETECTION AND PREVENTION OF ATTACKS USING ARTIFICIAL INTELLIGENCE IN THE BANKING SYSTEM

UDC 336.717
DOI: 10.36871/ek.up.p.r.2025.05.10.008

Authors

Anna A. Ilmushkina,
Russian Biotechnological University, Moscow, Russia
Veronika A. Danilova,
Russian State University of Tourism and Service, Moscow, Russia
Angela S. Yusupova,
Kadyrov Chechen State University, Grozny, Russia

Abstract

This article discusses modern methods of detecting and preventing cyberattacks in the banking system using artificial intelligence algorithms. The relevance of the topic is due to the sharp increase in the number of targeted attacks implemented using automated tools and machine learning technologies. An analytical review of domestic and foreign studies is conducted, including the work of Russian specialists in the field of graph analysis of transactions and protection of neural network models from adversarial influences. Key attack vectors are classified: from AI-enhanced phishing and malicious generative models to poisoning of training samples and insider threats using intelligent forecasting. The architecture of a multi-level cyber defense system is proposed, including anomaly detectors, autoencoders, GNN models and SIEM/SOAR solutions with real-time response.

Keywords

artificial intelligence, banking system, cyber threats, attack detection, attack prevention, anomaly detection