AI-Driven Malware Detection and Classification: A Systematic Review of Techniques and Effectiveness

In addition to classifying malware, it was further observed that malware analysis experts have developed new methodologies and strategies to assess the composition of malware samples by comparing their behavior and features to several known malware families. Thus, this study examined AI-driven malware detection and classification, using the systematic review of literature to understand the techniques and effectiveness of the AI systems. The study adopted the meta-synthesis research design. Meanwhile, the PRISMA chart was used for the selection of literature. There were identified inclusion and exclusion criteria that were outlined to determine the literature that are relevant to the study. Results showed that the techniques used in AI-driven malware detection and classification systems include deep learning techniques, machine learning techniques, and hybrid models. The findings showed that the AI-driven malware detection and classification system is highly effective in detecting and classifying malware. Findings of the study showed that the evaluation strategies for AI-driven malware detection and classification include standard metrics, benchmark datasets, experimental comparisons, and cross-validation. Results showed that the challenges associated with the use AI-enhanced systems to detect and classify malware include computational complexity, interpretability, dataset limitations, adversarial attacks, and real-time deployment constraints. The study concludes that AI-driven malware detection and classification systems have different techniques and they are highly effective. It was recommended that there is a need for continuous update of datasets to reflect new attack vectors.