Enhancing graduate students’ academic writing through AI, peer, and instructor based multi-source feedback model

This study examines the effects of a multi-source feedback model that combines artificial intelligence (AI), peer, and instructor feedback in the online academic writing process in higher education. The aim of the research is to reveal the effects of this hybrid model on students’ feedback literacy, academic writing self-efficacy, and satisfaction levels. A mixed-methods design was used in the research. Quantitative data were collected through pre-test–post-test applications, while qualitative data were gathered through reflection forms and focus group interviews. The participants in the study were students enrolled in two graduate-level “Educational Research Methods” courses. The first group consisted of 15 master’s students, while the second group consisted of 16 doctoral students. AI feedback system was developed using the ChatGPT API and provided students with criterion-based and immediate feedback. The findings indicated that improvements in academic writing self-efficacy and feedback literacy were observed following the implementation of the multi-source feedback model, particularly among master’s students. While students reported highest satisfaction with instructor feedback, their satisfaction with AI feedback increased over time. In conclusion, the integrated use of AI, peer, and instructor feedback appears to provide a promising pedagogical approach, which may support students’ self-regulation, creativity, and continuous learning skills.