Lin, Yen-Yu

Language Center, National Chinyi University of Technology, Taiwan

Abstract: This study investigated how 24 Taiwanese university freshmen, as English as a Foreign Language (EFL) learners, used various AI writing tools and their impact on language production and development. Through a mixed-methods approach including writing tasks, workshops, screen recordings, and surveys, the research found that machine translation led to longer texts, while paraphrasing tools improved lexical variety and syntactic complexity. Proofreading tools primarily reduced errors. However, a post-test without AI tools showed no significant improvements in most linguistic features, except for a decrease in word concreteness. Translation tools were the most frequently used, followed by grammar checkers. Students generally viewed AI tools positively. The findings suggest that while AI tools assist revision, the internalization of language features varies, and over-reliance on machine translation may affect autonomous writing. Educators should consider ethical AI integration and redesign assignments to encourage strategic AI use beyond basic translation and grammar correction, contributing to a nuanced understanding of EFL learners’ interaction with AI writing technologies.

Key words: AI-powered writing tools, machine translation, automated written corrective feedback, paraphrasing

Note: This is a pre-proof version and is subject to change during the editing process.