USE OF AI ALGORITHMS FOR DETECTION AND BLOCKING OF CYBER ATTACKS ON MILITARY INFORMATION SYSTEMS

Abstract

In the modern geopolitical context, accompanied by the intensification of cyber conflicts, the problem of ensuring the cybersecurity of military information systems is of particular relevance. Such systems are critically important for national defense, operational command of troops and the implementation of strategic decisions, which makes them the main targets of both mass and targeted cyberattacks. Traditional approaches to cyberdefense, focused on signature or heuristic analysis, cannot provide an adequate level of protection against the latest threats that are constantly changing, adapting and using artificial intelligence tools. In this context, the use of artificial intelligence (AI) algorithms, in particular machine and deep learning methods, reinforcement learning and natural language processing, opens up new horizons for proactive, adaptive and autonomous cyberdefense. The article considers approaches to the development, implementation and operation of intelligent systems for detecting and blocking attacks in military conditions.The capabilities of Next-Gen IDS systems built on the basis of convolutional neural networks (CNN) are analyzed, which allow detecting DDoS attacks, port scanning and malicious activity in real time with high accuracy. Special attention is paid to the issues of processing data from multiple sources within the Data Fusion approach, which is critically important in combat conditions, where information comes from sensors, drones, satellites, etc.A special emphasis is placed on the role of reinforcement learning in decision-making in fast-paced combat situations, as well as on the use of edge computing technologies that ensure the autonomy of systems in case of loss of communication with the center. The risks associated with attacks on artificial intelligence models themselves, such as data poisoning, adversarial input, model inversion, are analyzed, and methods for their neutralization are presented. The integration of Explainable AI (XAI) is investigated to increase trust in system decisions, ensure transparency and the possibility of operational control.It is shown that the implementation of AI in military information systems allows for the implementation of self-learning defense platforms that are able to adapt to changes in the threat environment, generate new signatures based on current data, and coordinate responses within a centralized command and information infrastructure. The paper emphasizes the importance of an interdisciplinary approach, involving specialists in the field of information security, military logistics, and intelligence, as well as the creation of a holistic cyber defense doctrine based on AI. Therefore, AI algorithms are considered not as an auxiliary tool, but as the foundation of the future architecture of sustainable, flexible, and intelligent defense of military IT systems.