Asian Journal of Research in Computer Science

This study investigates global trends in artificial intelligence (AI)-driven cybersecurity through a combined systematic literature review and bibliometric analysis of publications from 2015 to 31^st July, 2025. As digitalisation accelerates and cyberattacks become increasingly sophisticated, AI has emerged as a transformative tool for threat detection, prevention, and response. The review identifies four core domains where AI applications are most prominent: anomaly-based intrusion detection systems, automated malware analysis, phishing and social engineering prevention, and Security Orchestration, Automation, and Response (SOAR). In these areas, machine learning and deep learning techniques, particularly convolutional and recurrent neural networks, autoencoders, and transformer-based models demonstrate superior performance in detecting complex, evolving threats compared to traditional rule-based approaches. The bibliometric analysis reveals exponential growth in research output since 2015, with a sharp rise between 2021 and 2023, coinciding with breakthroughs in generative AI, deep learning, and the increased cyber risks linked to the COVID-19 digitalisation surge. Citation patterns highlight the growing applied relevance of post-2020 research, while thematic evolution indicates a shift toward adversarial AI, federated learning, and zero-trust architectures. Despite significant advances, challenges persist around explainability, governance, dual-use risks, and global disparities in research capacity. This study underscores AI’s central role in shaping the future of cybersecurity while emphasising the need for ethical frameworks and equitable global participation in technological adoption.