Enhancing vocabulary learning in EAP: investigating AI-supported spaced retrieval, student experiences, and perceptions

Abstract

Abstract Vocabulary development is essential to successful learning in English for Academic Purposes (EAP), yet effective approaches for supporting spaced retrieval practice remain a persistent challenge. This mixed-methods study investigated changes in first-year EAP students’ vocabulary knowledge and their perceptions following an AI-supported spaced retrieval intervention, using quantitative tests and qualitative interviews ( N = 110). Quantitatively, pre- and post-test results were compared using paired-samples t -tests. Following the intervention, vocabulary scores increased significantly ( p < 0.001, d = 1.31). Qualitatively, students’ vocabulary-learning experiences were examined through thematic analysis of semi-structured interviews. Participants reported high levels of cognitive engagement, suggesting that the AI-supported activities helped them understand word meanings, usage, and contextual examples. However, many students reported infrequent voluntary use of AI tools outside the classroom, indicating that their behavioural engagement remained limited. Students also highlighted practical challenges, including difficulty formulating effective prompts and concerns about the reliability of AI-generated responses. These results point to a discrepancy between students’ actual AI-use behaviours and their perceptions of AI’s benefits for learning. Based on these findings, the study proposes a repeated vocabulary exposure framework and argues that generative AI should be viewed as a pedagogical support tool whose effectiveness depends on learners’ AI literacy and careful instructional design. Students’ prompt-response exchanges with XIPU AI were not collected; therefore, findings concerning AI use are based on test results and interview accounts rather than direct analysis of interaction logs.