Application of AI in Cyberattack Detection: A Review

Area of Science:

• Cybersecurity and Artificial Intelligence
• Intrusion Detection Systems
• Machine Learning and Deep Learning Applications

Background:

• Cyber-physical systems face increasing security threats from advanced cyberattacks.
• Traditional signature-based Intrusion Detection Systems (IDS) struggle against novel and zero-day attacks.
• Artificial Intelligence (AI) presents scalable, accurate, and adaptive solutions for modern cyberattack detection.

Purpose of the Study:

• To comprehensively review recent advancements in AI-based cyberattack detection techniques.
• To evaluate the strengths, limitations, and performance of various AI approaches on benchmark datasets.
• To identify key challenges and future research directions in AI-driven cybersecurity.

Main Methods:

• Review of Machine Learning (ML), Deep Learning (DL), Reinforcement Learning (RL), and Federated Learning (FL).
• Exploration of emerging techniques: generative AI, neuro-symbolic AI, swarm intelligence, lightweight AI, and quantum computing.
• Analysis of AI-driven anomaly-based and hybrid detection methods versus traditional signature-based IDS.

Main Results:

• AI-driven methods, particularly anomaly-based and hybrid approaches, show improved detection rates for unknown and zero-day attacks.
• Key challenges identified include computational costs, data quality, privacy, and model interpretability, with Explainable AI (XAI) offering solutions.
• Lightweight AI and quantum computing show potential for resource-constrained environments and enhanced detection efficiency.

Conclusions:

• AI is crucial for developing robust, interpretable, and efficient cyberattack detection systems.
• Future research should focus on updated datasets, hybrid quantum-classical models, and optimized Federated Learning (FL) protocols.
• Continued innovation in AI is essential for securing complex digital environments against evolving cyber threats.