AI-Powered and Conventional Malware Detection Approaches: Challenges and Future Trends

Abstract

The continuous increase in users on the internet and online services provisioning such as shopping and banking causes many cyber criminals with a massive number of hackers sniffing users’ data. Malware is increasingly evolving, and the growth of worms, viruses, spyware, trojan horses, and other developments of malicious code requires improved detection techniques which are directed to the use of dynamic malware detection techniques. As a result, there is a growing need for anti-malware methods for protection of internet users’ privacy. Conventional mechanisms, such as behavior-based and signature-based approaches, have been broadly used but face great challenges compared to the growing malware threats. Artificial Intelligence AI-driven methods, leveraging Deep Learning (DL) and Machine Learning (ML), offer better adaptability and detection rates. This article serves as a survey, which investigates conventional and AI-powered malware detection techniques. Also, it offers a comparative analysis model, such as adaptability, accuracy, scalability, resource requirements, cost and more to signal their strengths and weaknesses. Finally, the study highlights future developments to mitigate the limitations of both approaches and improve overall cybersecurity resilience.