Abstract This study explores the role of ChatGPT feedback and teacher feedback on students' writing performance, self-regulation, and effort beliefs. The primary aim was to investigate whether these feedback types influence the aforementioned variables during a 45-day intervention period. Two groups of elementary-level EFL learners participated in the study: one receiving teacher feedback and the other receiving ChatGPT feedback. Pretests and posttests were conducted to assess writing proficiency, self-regulation, and effort beliefs, accompanied by four writing tasks during the intervention. An analytic grammar rubric and questionnaires were employed for data collection, analyzed using descriptive and inferential statistics. The results revealed no statistically significant improvement in writing performance for either group, although both groups showed slight, non-significant gains. Regarding self-regulation, neither feedback type induced meaningful changes over time, and the interaction between time and treatment was not significant. However, a significant difference between the groups was noted, indicating higher baseline self-regulation in the teacher feedback group. For effort beliefs, a significant time effect was observed, with both groups exhibiting a decline from pretest to posttest, though no significant differences were found between the feedback types. These findings suggest that while both ChatGPT and teacher feedback provide comparable results in supporting students' effort beliefs and self-regulation, neither appears to significantly enhance these aspects during the intervention period. The decline in effort beliefs highlights potential challenges in maintaining students' motivation and perceptions of effort during prolonged learning phases, irrespective of the feedback source. Further research is recommended to explore the underlying mechanisms and long-term implications of AI-driven feedback in EFL contexts. Key terms: Chat GPT feedback, writing proficiency, Effort beliefs, Self-regulation, Writing self-regulation