AI-Supported Language Learning for Neurodiverse Students in Low-Resource Contexts: Student and Teacher Perspectives and Its Impact on Learning Outcomes

This study examined the potential of AI-assisted tools to improve English language learning for neurodiverse students (with ADHD, dyslexia, or autism) in low-resource settings in Iran, considering student and teacher perspectives and students’ lan-guage-learning outcomes. The study used a convergent mixed-methods design, and 142 neurodiverse learners and 97 teachers participated through surveys, a 4-week ex-perimental study involving 30 learners (15 AI intervention, 15 controls), and semi-structured interviews with 15 learners, 10 teachers, and five parents. The out-comes were positive: learners stated that they enjoy adaptive features such as multi-modal input and gamification (M=4.2/5) and are motivated by them, and teachers found inclusivity to be important but perceived low confidence (M=2.7/5) because of the training gaps. The AI group showed substantial improvements in vocabulary (+16.3, d=1.21), reading comprehension (+13.3, d=1.05), and oral fluency (+9.2 wpm, d=0.89) compared to controls. Qualitative themes emphasized personalization as em-powerment, as well as obstacles such as infrastructural constraints, exam-based cur-ricula, and cultural cynicism. Recommendations were provided on the transformative power of AI in promoting equity and the need to train teachers and make changes in low-resource schools.