Enhancing EFL students’ reading comprehension through AI-assisted tools: Evidence from Islamic higher education institutions

Abstract

The study examined the use of AI-assisted reading teaching to improve EFL students’ reading comprehension and the students’ perception of the use of AI in conducting Extensive Reading exercises in an Islamic higher education setting. A quasi-experimental pre-test-post-test control group design was employed with 32 fourth-semester English education students. In the experimental group, 17 pupils used Google Gemini for reading tasks and 15 students in the control group received standard instruction. The intervention was four weeks long. Reading comprehension was measured using 20 item essay test . Students’ impression was measured using five point likert scale questionnaire . The experimental group’s mean score rose from 63.94 on the pretest to 82.65 on the posttest, whereas the control group’s score rose from 68.87 to 74.27. The gain scores differed significantly between the two groups (t(30) = 13.90, p <.001). ANCOVA revealed that after accounting for pre-test scores, the groups differed significantly in post-test performance (F(1,29) = 145.36, p <.001, partial η 2 =.834). Students also had positive Attitudes toward AI assisted reading with an overall mean of 4.39; the highest evaluations were for engagement (M=4.50) and usefulness (M=4.40). These results suggest that generative artificial intelligence may be a focused cognitive support in conjunction with active reading, critical evaluation, and student responsibility for their own learning. The study emphasizes the necessity of incorporating cognitive support, technology acceptance and learner autonomy in the application of AI in EFL reading education.