The effectiveness of using artificial intelligence in improving academic skills of school-aged students with mild intellectual disabilities in Saudi Arabia.

PMID 39631141 | DOI 10.1016/j.ridd.2024.104884 · Research in developmental disabilities · 2025

View on PubMed ↗

WHAT THIS STUDY ADDS: Teaching academic skills to students with intellectual disabilities can be a challenge for educators. This study explored the role of AI in the field of special education and provided a novel framework for educational integration of children with mild ID. Interestingly, it demonstrated that using AI-driven interventions could significantly improve academic skills and lead to better learning outcomes. This is crucial, as the holistic approach associated with the "whole child" model of education is prevalent due to the challenges students with ID face in social interactions typically involved in academic socialization. BACKGROUND: Intellectual disabilities (ID) are neurodevelopmental disorders that hinder learning, communication, and daily functioning. It negatively impacts learning of academic skills such as match and reading skills. Artificial intelligence (AI) may offer opportunities for these students to improve their academic performance. AIMS: To investigate the effectiveness of using AI in improving academic skills in a sample of seventy boys aged 9-12 with mild ID. They were enrolled into special education programs integrated into public schools in the Jeddah region of Saudi Arabia. METHODS: Participants were randomly assigned to an experimental and control group. The experimental group received ten 60-minute sessions twice weekly during five weeks, utilizing AI to personalize their academic skills activities. The control group received the educational program without the AI. All participants completed the Woodcock-Johnson IV-achievement test at pre-intervention, post-intervention, and one-month follow-up. RESULTS: The experimental group demonstrated significant and sustained improvements in academic performance across all measured domains compared to the control group, with effect sizes ranging from moderate (η² = 0.685) to large (η² = 0.921), underscoring the efficacy of the intervention at post-intervention. CONCLUSIONS AND IMPLICATIONS: This study sheds light on the promise of applying AI tools in special education to respond to distinctive needs experienced by students with mild ID. Future studies should investigate the long-term effects of such interventions and their broader applications across diverse educational contexts for inclusive learning.