Changes in Learners’ AI Literacy and Affective and Linguistic Perceptions in AI-Integrated Process Writing

초록

This study investigated changes in EFL learners’ self-reported AI literacy and their affective and linguistic perceptions following participation in an AI-integrated, process-oriented English writing course. A total of 115 Korean university students completed pre- and post-surveys. Closed-ended responses were analyzed using paired-samples t-tests, and open ended responses were examined through content analysis. The results indicated a significant increase only in students’ self reported understanding of AI. However, changes in usage, evaluation, and ethical awareness remained limited and inconsistent, indicating the need for targeted support to develop functional, evaluative, and ethical competencies. Affective results showed a significant reduction only in self-reported anxiety, whereas other affective variables did not remain statistically significant after Bonferroni correction. Learners’ linguistic perceptions showed generally positive but uneven changes, with significant increases in perceived sentence structure and organization, whereas other language-related areas, such as grammar and vocabulary use, showed only minor, non-significant gains. Open-ended responses indicated a shift in learners’ perceptions of AI from error correction to broader writing-process support. At the same time, concerns shifted from AI’s reliability and accuracy toward overdependence and reduced learner autonomy. Overall, the findings suggest the importance of structured AI integration, including ethical guidance, scaffolding, and focused instructional support, to promote balanced and responsible AI use in writing.

키워드