PROSPECTS FOR INTEGRATION OF ARTIFICIAL INTELLIGENCE IN CYBERSECURITY

Read full text Summarise with VIRAT
Authors: Артем Васильович АНТОНЕНКО, С. О. Востріков, Serhii Chechyk, Д. Я. Сольський
Cited by: 0
References: 14

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The article defines artificial intelligence (AI) as a system that models aspects of human intelligence based on computer technology, mathematics, computer science and philosophy, with the aim of imitating perception, understanding and interaction with the environment. In the context of cybersecurity, AI is integrated into intrusion detection systems (IDS), where it increases efficiency through automated data analysis, threat prediction and adaptive response. The materials emphasize the relevance of such integration, especially in the context of increasing complexity of cyberattacks, including AI-driven threats such as autonomous malware and social engineering. Key research focuses on models such as IntruDTree - an ML-based system based on decision trees, which ranks security features, minimizes computational complexity and achieves high accuracy in intrusion detection, outperforming traditional methods (naive Bayes, logistic regression, etc.). The neuromorphic approach combines deep learning (DL) with neuromorphic processors, using autoencoders (AE) for unsupervised learning, discrete vector factorization (DVF) for weight transformation, and simulation on chips like IBM True North, achieving 90.12% accuracy in detecting malicious packets and 81.31% in classifying attacks. This provides energy-efficient, real-time detection in high-traffic networks. Machine learning is applied to analyze malware, detect zero-day attacks, anomalies, and traffic threats, with a focus on datasets like ADFA-LD, which are in line with current technologies for evaluating IDS. Risks include adversarial attacks on ML algorithms, requiring preventive measures. The AZ Safe Hacker Assets Portal collects data from hacker forums for proactive CTI, analyzing ML assets for search, navigation, and code comparison. Preventing AI attacks involves social engineering, where vulnerabilities are human-driven, but AI helps with education and mitigation. Bayesian methods enable quantitative risk assessment, situational awareness, and automation. Research (Gartner, Trend Micro) predicts that AI agents will dominate attacks and defenses, with a focus on quantum security, explainable AI, and autonomous systems, requiring a balance between innovation and risk.

Fields of Study & Subjects

Network Security and Intrusion Detection Network Packet Processing and Optimization Adversarial Robustness in Machine Learning

Publication details

Journal / source
Herald of Khmelnytskyi National University Technical sciences
Volume
363
Issue
2
Pages
26–31
Publication date
26 March 2026
Language
EN