Artificial Intelligence (AI) has become a strategic enabler of modern cybersecurity by enhancing organizations’ ability to detect, predict, and respond to increasingly sophisticated cyber threats. However, the rapid advancement of AI has also empowered cyber adversaries to develop more intelligent attacks, including AI-generated phishing, deepfake-enabled fraud, adaptive malware, and adversarial machine learning techniques. This study presents a structured narrative review of recent literature to examine the evolving role of AI in cybersecurity from both defensive and offensive perspectives. The review synthesizes current evidence on AI-driven cyber defense, emerging AI-enabled threats, and the technical, ethical, and governance challenges associated with AI adoption. Building on these findings, the paper proposes an AI-Driven Adaptive Cybersecurity Framework (AI-ACF) that integrates threat intelligence, AI analytics, explainable decision support, and adaptive response to support resilient cybersecurity operations. The review highlights that while AI has significantly improved threat detection, predictive intelligence, and automated incident response, its effectiveness depends on high-quality data, robust governance, explainable models, and meaningful human oversight. The study contributes to the growing body of knowledge by providing a balanced synthesis of AI’s dual role in cybersecurity and offering a conceptual framework that can guide future research and organizational implementation. Future studies should empirically validate the proposed framework across diverse operational environments to assess its effectiveness in enhancing cyber resilience.

Keywords: Artificial Intelligence; Cybersecurity; Machine Learning; Threat Intelligence; Explainable Artificial Intelligence; Adaptive Cybersecurity; Cyber Resilience.