Employing ChatGPT to improve high school students’ writing skills by providing feedback on topic-specific writing tasks

Abstract

Translate Article

Take Notes

The integration of Artificial Intelligence (AI) technology in language teaching has emerged as a potential method to enhance students’ writing skills in today’s digital era. This study investigated the effectiveness of ChatGPT in improving the writing skills of 11th-grade high school students in an English as a Foreign Language (EFL) context. Employing a mixed-methods, quasi-experimental design, and the research involved 30 students from a state high school, divided into an experimental group (receiving ChatGPT feedback) and a control group (receiving teacher feedback). Data were collected through a pre-test, six consecutive writing tasks, a post-test, a questionnaire for the experimental group, and semi-structured interviews with eight experimental group participants. Quantitative analysis, comparing task scores and overall performance, revealed that the experimental group showed immediate, statistically significant improvements in earlier tasks compared to the control group. However, these inter-group differences converged by the post-test, with no statistically significant difference observed in overall writing scores between the groups at the study’s conclusion. The control group even outperformed the experimental group in one later task. Qualitative findings from interviews and questionnaires indicated overwhelmingly positive student perceptions of ChatGPT, highlighting its accessibility, speed, and non-judgmental nature. Students self-reported feeling more confident, safe, and less stressed when receiving AI-driven feedback. These results suggest that ChatGPT can effectively supplement traditional teaching methods by providing efficient, personalized feedback, particularly for foundational writing mechanics, fostering learner autonomy, and positively impacting student attitudes. The study supports the strategic integration of AI-driven feedback systems to complement existing teaching methods, offering scalable and personalized support to enhance student learning outcomes.