A PRISMA systematic review of NLP-powered assistive technologies supporting visually impaired students

Abstract

This systematic review critically examines two decades of research on Natural Language Processing based assistive technologies for visually impaired learners, with a focus on functional scope, implementation maturity, and sustainability implications. Guided by a PRISMA-based methodology, peer-reviewed studies were systematically analysed to identify dominant technological approaches, application contexts, and patterns of institutional adoption. This study analysis indicates substantial progress in speech recognition, text-to-speech systems, conversational interfaces, and multilingual processing, expanding technical access to educational content. However, the literature remains predominantly oriented toward feasibility and short-term usability evaluation. Evidence addressing longitudinal resilience, examination environments, institutional integration, and regulatory alignment is comparatively limited. Through a multidimensional sustainability lens encompassing technical, social, economic, and institutional considerations, three structural gaps are identified: the limited transition from access to learner autonomy; insufficient integration within educational governance frameworks; and the absence of sustainability-oriented assessment models beyond performance metrics. Software-centric solutions embedded within existing educational workflows demonstrate comparatively stronger potential for continuity than hardware-intensive or proprietary implementations. Overall, the review positions NLP-based assistive technologies as important enablers of sustainable inclusive education while clarifying the structural and contextual conditions required for their long-term adoption and systemic impact.