Human vs ChatGPT Feedback in EFL Writing: A Comparative Study of Error Correction and Learner Uptake Among Saudi University Students

Abstract

The rapid spread of artificial intelligence (AI) in language education has led teachers to use this tool in teaching and pedagogical practices across several disciplines. In English as a foreign language (EFL) context, various AI tools are being applied in teaching practices. ChatGPT is one of the most popular tools currently used by both teachers and students. This research compares human feedback and ChatGPT feedback as two methods used in Saudi EFL settings. The study adopted a quasi-experimental design. The learner participants were divided into two groups: a human feedback group and a ChatGPT feedback group. The research focused on different writing concerns addressed by the two methods of feedback. These feedback notes include grammar, vocabulary, organization, and discourse. Also, the study focused on the different responses of the students to the two types of feedback.  The results revealed that ChatGPT feedback provided instant feedback on the surface level of the language structures. Human feedback, however, treated higher levels of contextual writing concerns. The study also focused on the students' rates of response to the two types of feedback. The results showed that the ChatGPT feedback group revised their grammatical mistakes rapidly. However, the students in the human feedback group spent a longer time dealing with contextual issues. Overall, the main recommendation of this research is that both types of feedback are helpful in improving EFL writing.