AI-Based Cyber Threat Detection in Modern Software Systems

Abstract

The high-paced development of modern software systems, which is described by cloud computing, microservices, the Internet of Things (IoT), and distributed architecture, has greatly increased the area of cyber-attacks, making the traditional signature-based security mechanisms less effective. To overcome these difficulties, artificial intelligence (AI) has been proposed as a groundbreaking strategy to detect cyber threats, which provides predictive, adaptive, and autonomous security features. This review article critically examines AI-based methods of cyber threat detection, which is used in modern software systems. It reviews the literature on machine learning, deep learning, and hybrid AI applications in malware, intrusion attempt, phishing attack, insider threats, and zero-day vulnerability detection. In addition, the review presents information related to data source, feature engineering tactics, and real-time detection models that foster intelligent threat detection in the cloud, IoT, and enterprise settings. The major performance indicators, implementation difficulties, and scalability challenges related to the AI-based security systems are discussed with utmost importance. There is also consideration of ethical concerns such as data privacy, the explainability of the model, and adversarial attacks on the AI systems. This review will additionally illuminate the way forward in developing resilient, explainable, and robust AI-enabled cyber defense systems by integrating current developments and highlights unresolved research gaps that will shape the future of software systems based on next-generation software systems.

Keywords: Artificial Intelligence, Cyber Threat Detection, Software Security, Machine Learning, Deep Learning, Intrusion Detection Systems, Malware Detection, Zero-Day Attacks, Cloud Computing Security, Internet of Things (IoT) Security, Adversarial Attacks, Explainable AI.