Keystroke Dynamics Authentication: A Comprehensive AI-Driven Review

Keystroke dynamics (KD), a behavioral biometric method analyzing typing patterns, provides reliable user authentication/identification. This systematic review explores the role of artificial intelligence (AI) in enhancing KD-based authentication. Based on 100 studies from IEEE Xplore, Scopus, PubMed, and ACM Digital Library (2015–March 2025), the analysis demonstrates how neural networks, support vector machines (SVMs), hybrid models, and deep learning architectures have significantly improved accuracy and robustness. Notable advancements include LSTM-based systems achieving near-perfect precision and hybrid CNN-SVM frameworks balancing speed and performance. However, challenges persist, including privacy risks from typing pattern re-identification, vulnerabilities to adversarial attacks, and deployment on resource-constrained devices. By synthesizing these findings, this study highlights key limitations and proposes future research directions to develop robust, ethical, and accessible KD authentication systems.