A Systematic Review and Taxonomy of Ransomware Detection Based on Artificial Intelligence Algorithms

Abstract

The escalating prevalence of ransomware attacks poses a significant risk to digital infrastructures, data integrity, and essential services worldwide. Traditional signature-based detection methods often struggle to keep pace with the evolving landscape of ransomware variants and stealthy attack techniques. Artificial Intelligence (AI) offers a promising solution, leveraging sophisticated features such as bytecodes, opcodes, API calls, and behavioral analysis to enhance ransomware detection capabilities. This review delves into the latest advancements in ransomware prevention and detection techniques, categorizing ransomware types and examining the intricate lifecycle of ransomware attacks. A comprehensive taxonomy of ransomware analysis techniques, machine/deep learning methods, and feature selection techniques is presented, considering the diverse operating systems targeted by these malicious threats. By conducting a thorough analysis of recent research articles in this field, we identify the key challenges faced by academics and the research community in mitigating the ransomware threat. Moreover, we explore potential future research directions to address these challenges and develop more effective countermeasures.