<![CDATA[This study investigates the impact of AI-enhanced multimodal reading on EFL learners’ comprehension and engagement in an Indonesian madrasah context. Employing an explanatory sequential mixed-methods design, the research combined a pre-experimental one-group pretest–posttest design with semi-structured interviews. Thirty eleventh-grade students at MAS Al Falah, Tulang Bawang, participated in the quantitative phase, while six students were purposively selected for the qualitative phase. The intervention, conducted over four weeks, integrated DALL·E-generated visuals into reading tasks covering topics such as Indonesian environmental figures, digital literacies, and personal identity. Quantitative findings revealed a significant improvement in learners’ comprehension, with mean scores rising from 75.07 (pretest) to 81.73 (posttest), and an N-Gain of 0.269 indicating low-to-moderate progress. Qualitative analysis further explained these gains, showing that students attributed their improvement to visual scaffolding, heightened motivation, stronger identity connection, and collaborative meaning-making, although some expressed challenges in interpreting AI-generated images. The integration of both strands through joint display confirmed that cognitive, affective, and sociocultural dimensions collectively shaped learning outcomes. The study concludes that AI-enhanced multimodal reading can enrich EFL pedagogy by promoting comprehension and engagement, while also highlighting the need for scaffolding and digital literacy development. Future research should extend to longitudinal designs, larger populations, and comparative analyses of different AI tools to sustain and generalize the benefits of multimodal learning. ]]>
