Artificial Intelligence (AI) Applications in Personalized Learning for Autistic Students: A Systematic Literature Review

Abstract

Personalized learning is a significant approach in the education of autistic students due to the diversity of their learning profiles, particularly in terms of social communication, interaction, and sensory sensitivity. However, the optimal implementation of personalized learning in classrooms continues to face multiple constraints, including teacher workload and challenges in providing interventions that are truly individualized on a consistent basis. While Artificial Intelligence (AI) technology is seen as having the potential to support this process through more adaptive applications, existing literature on its implementation remains varied, fragmented, and highlights research gaps. Therefore, this study was conducted as a Systematic Literature Review (SLR) with the objectives of identifying the types of AI instruments commonly used, assessing their impact on social communication, emotion management, and student engagement, and examining the challenges and research gaps in the use of AI within special education. The study employed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and searched six major databases: Scopus, Web of Science (WoS), ERIC, IEEE Xplore, ACM Digital Library, and Google Scholar. From a total of 226 identified articles, 24 publications from 2019 to 2026 met the inclusion criteria and were analyzed thematically. The findings indicate that the most frequently used AI instruments include Intelligent Tutoring Systems (ITS), social robots such as NAO, adaptive learning systems, and emotion recognition technologies. The use of these technologies has been shown to positively impact student focus and engagement, support social communication, and assist in emotion management through personalized activities. Nevertheless, the implementation of AI interventions also faces major challenges, including technical constraints, high training demands for teachers, and concerns regarding data privacy and ethics. Accordingly, this study recommends that AI should be employed as a pedagogical support tool, integrated with human interaction and professional teacher training. The findings are expected to provide meaningful contributions to inclusive pedagogical practices and to the development of safer, more equitable educational technology policies that respect and accommodate the neurodiversity of autistic students.