The Effect of Self-Regulated Use of Chat GPT on the Paragraph Writing Skills and Self-Efficacy of Iranian EFL Learners

Document Type : Original Article

Authors

Dept. of English Language and Literature, Faculty of Literature and Languages, Arak University, Arak

Abstract

As technology has advanced rapidly, various AI tools have been integrated into educational contexts. This study investigates the effects of self-regulated use of ChatGPT on paragraph-writing skills and self-efficacy among Iranian EFL learners. To this end, 64 university students were assigned to experimental and control groups in two intact writing classes over a twelve-week intervention. The experimental group received training on self-regulated learning (SRL) strategies based on Zimmerman’s (2002) established model. Key strategies covered included: goal setting, self-monitoring, self-evaluation and reflection, and strategic resource management. They also utilized ChatGPT for writing tasks, whereas the control group followed traditional instruction. Pre- and post-tests of compare-contrast and cause-effect paragraph writing were scored for organization, logical development, grammar, mechanics, and style. Self-efficacy was measured using a validated questionnaire. The results indicated significant differences in the post-test mean scores of the experimental and control groups on the compare-and-contrast and cause-and-effect sub-sections, both of which represent large effect sizes. The findings suggest that integrating AI-driven tools facilitates self-regulated learning, enhances writing proficiency, and strengthens learners’ writing confidence. These results have implications for educational policy and curriculum development, underscoring that adopting AI tools complements traditional writing instruction by facilitating a more personalized learning experience.

Keywords

Main Subjects