The role of artificial intelligence in malicious code analysis

Abstract

this paper examines the role of modern artificial intelligence (AI) methods in the detection and analysis of malicious saroftwe. The main approaches are reviewed-static, dynamic, and hybrid analysis based on machine learning (ML) and deep learning (DL) algorithms-together with the practical problems that limit their deployment: code obfuscation, adversarial attacks on models, dataset scarcity and drift, limited interpretability, and detection of zero-day threats. It is shown that combining classical signature-based methods with AI-driven models improves detection accuracy, reduces analyst workload, and shortens response time to emerging threats. Directions for further research, including explainable AI and adversarially robust models, are outlined.