The Use of Artificial Intelligence for Cyber Threat Detection: A Systematic Literature Review of Research Methods, Accuracy, and Gaps

Muhammad Distian Andi Hermawan (Management Study Program, APRIN College of Economics (STIE APRIN), Palembang, Indonesia)

Journal of Computing and Smart Ecosystems
Vol. 1 No. 2 (2025): J-CaSE

This study presents a comprehensive Systematic Literature Review (SLR) on the use of Artificial Intelligence (AI) for cyber threat detection, focusing on methods, accuracy levels, and research gaps from the last five years. A total of 47 eligible studies were analyzed using the PRISMA framework. The findings show that deep learning has become the dominant approach, outperforming traditional machine learning in identifying complex threats such as DDoS, zero-day attacks, and advanced malware. Hybrid models also demonstrate high accuracy, exceeding 95% in several datasets. However, significant gaps remain, including limited real-time evaluations, outdated public datasets, insufficient research on explainable AI, and the lack of adversarial defense mechanisms. This review emphasizes the need for more robust, interpretable, and adaptive AI-based security systems to address evolving cyber threats effectively. The results provide essential insights and guidance for future research in AI-driven cybersecurity.