Enhancing cybersecurity in online learning environments: a systematic review of AI-driven approaches

The rapid shift to online learning environments (OLEs), accelerated by the COVID-19 pandemic, expanded the attack surface of educational systems and introduced vulnerabilities that traditional security mechanisms were not designed to handle. In this context, artificial intelligence (AI) has attracted growing interest as a response, offering proactive, automated defenses that rule-based approaches struggle to replicate. This paper presents a systematic review of AI-driven cybersecurity in OLEs using a mixed-methods approach that combines thematic, descriptive and meta-analytic analyzes in line with PRISMA guidelines. The review investigates the CIAAN framework alongside core security dimensions (threats, vulnerabilities and risks) and evaluates mitigation strategies across traditional, AI-based and hybrid approaches. It further provides an in-depth analysis of AI-specific components, including architectures, performance metrics and datasets, while examining adoption challenges. The findings demonstrate that across the three cybersecurity tasks studied (face authentication, intelligent proctoring and network intrusion detection), AI-based approaches consistently outperformed conventional methods. However, AI adoption within OLEs remains limited, with infrastructure constraints, usability challenges and privacy concerns as the primary obstacles. The review draws on these findings to offer practical guidance for institutions seeking to integrate AI into their cybersecurity frameworks.