Explainable AI Driven Frameworks for Real Time Cyber Threat Detection and Response: A Systematic Review

The digitization of an ever-developing network of cybersecurity threats has weighted traditional packet filter rule-based firewalls and signature based intrusion detection systems towards being less effective against zero day vulnerabilities, ransomware, polymorphic malware, and Advanced Persistent Threat (APTs) attacks. Emerging technologies like Artificial Intelligence (AI) and Machine Learning (ML) have potential to be game changers for proactive and adaptive defence of cyberspace. This review investigates AI powered cybersecurity systems: machine learning algorithms, deep learning architectures, Reinforcement Learning, and Explainable Artificial Intelligence (XAI). Also investigated will be applications within Cloud Computing, Healthcare, Maritime Systems and National Critical Infrastructure. Further discussions will include feature engineering, Transformer based Zero Shot Learning (ZSL), SOAR integration, blockchain-based forensic integrity, adversarial AI and quantum computing concerns. The results demonstrated positive trends through AI powered cybersecurity systems improving detection Accuracy, lowering False Positives, reducing incident response time and enhancing Autonomous Defence capabilities. However, challenges around Explainability, Ethical Governance, Scalability of Computation, and Adversarial Resistance are expressed as possible major concerns. A conclusion point will outlining future research directions to develop, build and maintain resilient, transparent and quantum resistant Cybersecurity Ecosystems.