AI-DRIVEN FEEDBACK FOR VOCABULARY DEVELOPMENT: A BEFORE-AFTER ANALYSIS IN INDONESIAN TERTIARY EFL WRITING

This study aims to analyze the effectiveness of AI-driven feedback in enhancing vocabulary development among Indonesian tertiary EFL (English as a Foreign Language) students. Using a before-after research design, the study involved 60 undergraduate English majors from a public university in Indonesia. Participants were asked to write two academic essays—before and after receiving automated feedback from AI-based platforms such as Grammarly and ChatGPT. The analysis focused on improvements in lexical complexity, vocabulary range, and appropriate collocation use. Quantitative data were examined through paired-sample t-tests, supported by qualitative analysis of the most influential feedback types contributing to lexical improvement. The findings revealed a significant increase in students’ vocabulary range and depth after the application of AI-driven feedback, with a significance level of p < 0.05. Moreover, students reported that AI feedback offered more personalized, immediate, and exploratory learning compared to manual correction from lecturers. However, the study also identified limitations related to students’ overreliance on automated suggestions without deeper linguistic understanding. The results suggest that integrating AI-driven feedback into EFL academic writing can serve as an effective strategy to foster lexical competence, provided it is complemented by appropriate pedagogical guidance from instructors.