Abstract
The complexity and increasing rate of cyberattacks have become a formidable hurdle to people, institutions, and infrastructure vital sectors in all continents. Conventional signature-based defenses to cybersecurity threats, which are usually backreactions by nature, can hardly keep up with trends in threats that are constantly shifting. Artificial Intelligence (AI), including Machine Learning (ML) and Deep Learning (DL), is the new paradigm enabling completely new opportunities to enhance cybersecurity mechanisms. This paper is an in-depth overview of how AI can be used to enhance cybersecurity, the various uses and applications, techniques, and all other future possibilities. We also get into the various aspects of AI being used in cybersecurity, which includes intrusion detection, malware analysis, security orchestration, threat intelligence, and vulnerability management.
Next, we address the challenges and limitations of the implementation of AI in cybersecurity, that is, the data privacy problems, adversarial trained AI attacks, and explainability. Lastly, we point out the emergent lines of future research with the most potential, such as blockchain + AI, further development of Explainable AI (XAI), and combining human interaction with AI. The review is intended to be a useful tool for both researchers, practitioners, and policymakers in order to learn more about AI and how it can be used to develop more proactive and resilient cybersecurity.
