AI-Enhanced Science Learning for Primary Students With Autism Spectrum Disorder: A Quasi-Experimental Study With...

This study examines the impact of Artificial Intelligence (AI)-enhanced instructional tools on science learning, engagement, and social interaction among primary school students with autism spectrum disorder (ASD). The research addresses a gap in the literature by exploring how adaptive, multisensory technologies can support conceptual understanding of "Light and Color" and foster inclusive learning experiences in special education settings.

A quasi-experimental pre-test/post-test design was implemented with 34 students with ASD, divided into an experimental group and a control group receiving traditional teaching. Quantitative data were collected through science knowledge tests, teacher-rated engagement scales, and AI-generated learning analytics. Qualitative data were obtained through structured observations and teacher reflective journals. Statistical analyses included paired t-tests, ANOVA, MANOVA, and effect size calculations, while qualitative data underwent thematic analysis.

Students in the AI-enhanced group demonstrated significantly greater improvements in conceptual understanding compared to the control group, with large effect sizes across most test items. Engagement and motivation ratings were also higher for the experimental group. Qualitative observations revealed increased curiosity, sustained attention, collaborative behaviors, emotional regulation, and autonomy when interacting with the robot and VR environments.

AI-enhanced, multisensory instruction substantially benefits neurodiverse learners by promoting deeper conceptual learning, supporting emotional and behavioral engagement, and enhancing inclusive participation. Findings highlight the pedagogical value of integrating adaptive AI tools into science education for students with ASD and suggest that AI can serve as an effective co-mediator alongside teachers in special education contexts.