<![CDATA[Artificial Intelligence (AI) is increasingly transforming education through personalization, automation, and data-informed decision-making, yet its application in correctional education remains underexplored. This study examines the implementation of AI-enabled e-learning in correctional centres worldwide, emphasizing secure, adaptive, and ethical deployment. A systematic literature review following the PRISMA framework analysed peer-reviewed studies published between 2019 and 2025. No empirical studies were identified that evaluated fully AI-enabled learning systems on-site in correctional settings. Instead, the available evidence focuses on AI-adjacent educational technologies, secure offline platforms, and adaptive learning management systems. Findings indicate that digital education is increasingly feasible in correctional environments, although implementation is constrained by security protocols, limited infrastructure, and resource availability. Reported benefits include improved learner motivation, autonomy, and digital literacy, while evidence linking higher education to reduced recidivism remains cautious. AI-driven adaptive and personalized learning therefore remains largely aspirational. Significant ethical and legal gaps also persist, particularly regarding algorithmic bias, privacy, and data governance. Correctional education is consequently at a transitional stage requiring stronger regulation, participatory design, and longitudinal research to ensure that AI supports rehabilitation without reinforcing existing inequalities.]]>
