Beyond the Textbook: Exploring AI-Assisted Materials Development as an Emerging Practice in EFL Instruction

Artificial intelligence (AI) is increasingly used to support instructional materials development in English as a Foreign Language (EFL) context. This explanatory sequential mixed-methods study examined whether AI-generated instructional materials influenced EFL learners’ academic performance and explored teachers’ perceptions of their usability and pedagogical value. A quasi-experimental pretest-post-test control-group design was implemented over 16 weeks in an intermediate intensive English course at a higher education institution in Oman. Eight intact classes (N = 226) were randomly assigned at class level to the experimental condition (n = 110) or control condition (n = 116). The experimental group was taught using materials generated with Twee and adapted by teachers, whereas the control group received conventionally prepared materials. Quantitative data were collected through institutional intermediate-level English achievement tests in listening, reading, and writing, and qualitative data were obtained through semi-structured interviews with the four lecturers who taught the experimental classes. MANCOVA results showed a significant overall group effect on the combined post-test scores after controlling for pretest performance, Wilks’ ? = .916, F(3, 219) = 6.72, p < .001. After Bonferroni correction across the three skill-specific follow-up tests, writing remained significant, whereas listening and reading did not. A supplementary ANCOVA indicated a significant group difference in overall performance. Teachers perceived the AI-generated materials as accessible and pedagogically useful, while stressing the need for careful review and adaptation before classroom use. The study highlights the value of teacher-mediated AI-assisted materials development for technology-enhanced EFL instruction.