CHATGPT’S AND HUMANS’ WRITTEN CORRECTIVE FEEDBACK IN EFL/ESL WRITING: A SYSTEMATIC REVIEW
Abstract
Written corrective feedback (WCF) is critical in EFL/ESL writing classrooms. This study aims to review existing research studies on the quality of WCF generated by ChatGPT and humans and to examine how WCF generated by ChatGPT can support WCF that humans give to students’ written work in higher education EFL/ESL writing contexts. To meet these goals, the researcher examined 23 studies published in 18 peer-reviewed journals from 2023-2024. The analysis results highlighted potential strengths and weaknesses of WCF generated by ChatGPT and humans. WCF generated by ChatGPT could review various aspects of writing and provide extensive descriptive feedback, yet it sometimes might experience system fatigue that affects the WCF quality. ChatGPT could also provide inaccurate WCF, give WCF that lacked human nuances, and were greatly influenced by how prompts were presented to it. Meanwhile, humans could provide aspects of WCF not accounted for ChatGPT, but their WCF might involve subjective judgment and be unclear. With these findings in mind, the researcher identified three phases of how ChatGPT can work harmoniously with humans to provide quality WCF for students in writing classrooms. Those phases were making various efforts to enhance WCF generated by ChatGPT in students’ writing in support of the feedback provided by lecturers, enhancing students’ in-depth understanding that not all ChatGPT feedback is accurate, and combining ChatGPT-based WCF, specifically on earlier drafts of students’ writing. Pedagogical implications, which highlighted the essence of working synergistically with ChatGPT and mastering prompts to generate the most helpful WCF, for EFL/ESL lecturers in writing classrooms who wish to support their feedback with ChatGPT to provide quality WCF to their students’ written work were presented. The researcher then proposed recommendations for future systematic review studies, expanding the research contexts, the scope of reviewed materials, and the research method of the reviewed articles, to address gaps and expand on the insights presented in this study.
Keywords
written corrective feedback, ChatGPT, EFL, ESL, writing classrooms
References
- Abas, I. H., & Abd Aziz, N. H. (2018). Model of the writing process and strategies of EFL proficient student writers: A case study of Indonesian learners. Pertanika Journal of Social Sciences and Humanities, 26(3), 1815–1842.
- Ajlouni, A. O., Abd-Alkareem Fatima, W., & Almahaireh, A. S. (2023). Students’ attitudes towards using chatGPT as a learning tool: The case of the University of Jordan. International Journal of Interactive Mobile Technologies, 17(18), 99–117. https://doi.org/https://doi.org/10.3991/ijim.v17i18.41753
- Alberth. (2023). The use of chatgpt in academic writing: A blessing or a curse in disguise? TEFLIN Journal, 34(2), 337–352. https://doi.org/10.15639/teflinjournal.v34i2/337-352
- Algaraady, J., & Mahyoob, M. (2023). ChatGPT’s capabilities in spotting and analyzing writing errors experienced by EFL learners. Arab World English Journal, 9, 3–17. https://doi.org/https://dx.doi. org/10.24093/awej/call9.1
- Alkamel, M. A. A., & Alwagieh, N. A. S. (2024). Utilizing an adaptable artificial intelligence writing tool (ChatGPT) to enhance academic writing skills among Yemeni university EFL students. Social Sciences and Humanities Open, 10, 1–9. https://doi.org/10.1016/j.ssaho.2024.101095