Integrating synergistic learning and ability grouping with AI-generated multi-tiered tasks to enhance reading attitudes and outcomes

Many EFL classrooms include students with diverse proficiency levels, complicating efforts to improve reading comprehension through one-size-fits-all instruction. This study addressed that challenge by examining how integrating AI-generated multi-tiered reading tasks, proficiency-based ability grouping, and a synergistic collaborative learning environment affects the reading comprehension and engagement of university EFL students. In a nine-week quasi-experimental mixed-methods design, 87 female first-year university students were divided into proficiency groups based on the CEFR-J (A2 - Group 1, B1.1 - Group 2, and B1.2 - Group 3), and exposed to AI-generated reading materials tailored to each level, ensuring content that was appropriately challenging for each group. The synergistic classroom setting enabled peer scaffolding and shared learning experiences by encouraging synergy both within and between groups of students with similar ability levels. Pre- and post-intervention reading tests (IELTS-aligned) showed significant gains in reading comprehension across all proficiency levels, with the lowest level (A2) making robust progress. Likert-scale questionnaires revealed that learners were highly engaged and satisfied, and expressed positive sentiments toward the differentiated, collaborative method. Conversely, the qualitative classroom observations and semi-structured interviews indicated increased learner autonomy and emotional involvement, as students exhibited greater confidence and investment in their reading. The integrated strategy of AI-assisted differentiation and synergistic ability grouping significantly enhanced reading comprehension while promoting increased autonomy and personalization. These findings underscore the model's potential as a scalable solution for EFL environments with varied learner requirements, utilizing technology to customize learning and foster peer-assisted development.