Artificial intelligence for engaging neurodiverse learners in K-12 education: a systematic review

While artificial intelligence (AI) has been widely explored in educational research, limited studies have specifically examined its intentional use to support K-12 neurodiverse learners. Existing literature often conflates clinical diagnosis, assistive technology, and instructional design, obscuring AI’s distinct pedagogical contributions. This qualitative systematic review addresses this gap by exploring how AI tools are used to engage and support neurodiverse learners across formal and informal K-12 learning environments. A comprehensive search was conducted in Scopus and the Web of Science using Boolean operators, following PRISMA 2020 guidelines, yielding 3032 initial articles, with 15 studies meeting inclusion criteria for analysis. Guided by the PICOTS framework, a deductive thematic analysis was employed to identify themes across AI system features, learner interactions, and pedagogical integration. The three-paradigm model of AI integration – AI-directed, AI-supported, and AI-empowered, and the TPACK frameworks informed the synthesis. Findings reveal that while AI shows promise in personalizing learning for students with neurodiverse profiles, most implementations remain teacher-directed, with limited movement toward learner empowerment. Key considerations for effective AI adoption include pedagogical alignment, explainability, teacher training, and ethical use. This review offers suggestions for future AI integration, emphasizing adaptive, meaningful, and inclusive practices tailored to neurodiverse learners.