Characteristics of Grammatical Feedback Between ChatGPT and Lecturer Corrections on Students' Writing Tasks: A Qualitative Comparative Study
DOI:
https://doi.org/10.59944/jipsi.v4i4.2190Keywords:
ChatGPT, Written Corrective Feedback (WCF), EFL writing, lecturer feedback, Automated Writing Evaluation (AWE)Abstract
This study investigates the linguistic characteristics and pedagogical depth of written corrective feedback (WCF) provided by a human lecturer versus ChatGPT-4 on undergraduate EFL writing tasks. Employing a descriptive-comparative content analysis method, the study evaluated five student argumentative essays alongside corresponding manual lecturer annotations and standardized ChatGPT feedback logs. The findings indicated a substantial qualitative disparity, the lecturer’s feedback functioned primarily as direct corrective feedback, relying on visual shorthand markup (strikethroughs, insertion arrows) to deliver localized, surface-level fixes without metalinguistic explanations. Conversely, ChatGPT functioned as an interactive formative writing tutor, offering detailed metalinguistic explanations, full-sentence reconstructions, formal vocabulary substitutions, and macro-level discourse guidance, including active voice recommendations, redundancy removal, and transitional flow improvements. The study concluded that while ChatGPT excels in comprehensive, line-by-line syntactic and stylistic feedback, it lacks the human empathy and contextual intuition of an instructor. Consequently, the paper advocates for a blended, hybrid feedback model in EFL pedagogy: students utilize ChatGPT as an automated pre-submission screening tool for lower-order mechanical corrections, allowing human instructors to focus their feedback on higher-order priorities such as argument coherence, critical thinking, and authentic voice.
References
AlGhamdi, A. (2024). Generative AI in the EFL classroom: Exploring student reflection logs on ChatGPT-assisted writing revisions. Journal of Language and Linguistic Studies, 20(1), 112–126. tandfonline.com
Asadi, A., Ebadi, S., & Mohammadi, M. (2025). The pedagogy of hybrid feedback models: Integrating artificial intelligence with instructor guidance in EFL context. Computer Assisted Language Learning, 38(2), 204–225. researchgate.net
Athanassopoulos, S., Manoli, P., Gouvi, M., Lavidas, K., & Komis, V. (2023). The use of ChatGPT as a learning tool to improve foreign language writing in a multilingual and multicultural classroom. International Journal of Educational Technology in Higher Education, 20(1), 45–62.
Barrot, J. S. (2026). Mapping thematic evolution and pedagogical strategies of Generative AI in EFL writing. ELT Journal, 80(1), 12–25.
Dararat, K. (2026). The impact of ChatGPT on academic writing skills and knowledge: An investigation of its use in argumentative essays. Journal of English Language Teaching, 15(1), 89–104.
Escalante, J., Flores, M., & Rodriguez, L. (2023). Evaluating AI-generated feedback in language acquisition for beginner English learners. Computer Assisted Language Learning, 36(4), 512–530.
Funabashi, R. (2024). Generative AI in academic writing: English for academic purposes practitioners' perceptions and teaching practice (Dissertation). BALEAP MA Dissertation Award Winners. baleap.org
Guo, K., & Wang, Y. (2023). Investigating the application of ChatGPT-4 in EFL writing instruction: A multi-dimensional feedback perspective. Interactive Learning Environments, 31(8), 4412–4429.
Hakiki, M., Rahmawati, A., & Putera, D. (2023). Integrating ChatGPT into university technical education to improve academic performance. Journal of Vocational Education and Studies, 6(2), 201–215.
Han, Y., & Li, X. (2024). Observational analysis of classroom interactions during AI-mediated English writing feedback sessions. Language Teaching Research, 28(3), 302–321. tandfonline.com
Ju, H. (2023). The dilemma of generative AI assistance: Analyzing the decline in learning efficiency in English writing tasks. Language Learning & Technology, 27(3), 104–121.
Kim, N. Y., Shin, S., & Lee, H. (2024). Complementary tools or competitors? ChatGPT-generated and instructor-delivered formative feedback in Korean undergraduate EFL writing courses. Journal of Educational Computing Research, 62(3), 315–334.
Kohnke, L., Moorhouse, B. L., & Zou, D. (2023). ChatGPT for language teaching and learning: Opportunities and challenges. TESOL Quarterly, 57(2), 537–540.
Kusuma, I. P. I., Wardana, I. K., & Suputra, G. A. (2024). Immediate feedback and paraphrasing assistance: The role of ChatGPT in secondary writing instruction. Journal of Digital Learning and Education, 4(1), 45–58.
Lee, S., & Jeong, H. (2026). Large language models as interactive partners: Correcting syntactic errors and vocabulary expansion in L2 writing. System, 118, 103–118.
Lo, C. K. (2023). What is the role of ChatGPT in education? A systematic review of early research. Education Sciences, 13(4), 410–425.
Mahapatra, S. (2024). Peer discussions vs AI evaluation: Social and collaborative elements of EFL writing feedback. TESOL Journal, 15(2), 144–159. tandfonline.com
Memarian, B., & Doleck, T. (2023). ChatGPT in education: Methods, evaluation, and automated feedback system capabilities. Computers and Education: Artificial Intelligence, 5, 100–114.
Mizumoto, A., & Eguchi, M. (2023). Exploring the accuracy of GPT models in corpus-based grammatical error correction for L2 learners. English for Specific Purposes, 72, 85–98. tandfonline.com
Mun, C. Y. (2024). EFL college learners' AI-assisted writing feedback via ChatGPT vs. peer feedback. STEM Journal, 25(2), 26–39.
Ng, D. T. K., Liang, J., & Chu, S. K. W. (2024). Utilizing AI chatbots (SRLbot) to enhance student engagement and reduce language anxiety in educational contexts. Computers & Education, 210, 104–119.
Nur, M., & Arifin, S. (2025). Peran ChatGPT dalam pembelajaran menulis bahasa Inggris bagi mahasiswa Ekonomi Syariah. Journal of Soshum (J-SES), 12(2), 143–155.
Pratama, R., & Wibowo, A. (2026). Efektivitas ChatGPT dalam pembelajaran bahasa Inggris untuk mendorong self-regulated learning di perguruan tinggi. Jurnal P4I: Learning and Instruction, 14(1), 77–89.
Shi, L., Wang, Z., & Zhang, X. (2026). LLM versus traditional AWE: Comparing ChatGPT and traditional automated writing evaluation systems in an 11-week intervention. MDPI Education Sciences, 16(8), 1196–1211.
Wibowo, S., & Hidayat, R. (2025). Integrasi Artificial Intelligence dalam pembelajaran bahasa Inggris di perguruan tinggi: Peluang pedagogis dan tantangan. Jurnal Inspirasi Modern (JIM), 3(2), 88–99.
