AI-Generated English Feedback versus Human Feedback: Comparative Effects on Writing Accuracy and Creativity

A. S. M. Saiful Islam, MD Ashickur Rahman, Musadderul Hoque & S M Saadat Hossain

Jan-Feb-Mar

Abstract
The paper is mixed-methods research that seeks to compare the various effects of AI-generated feedback and traditional feedback provided by human teachers on writing accuracy, creativity, and self-correction behavior in students. The main aim was to identify whether AI-based grammar correction software can be effective in improving the writing abilities of learners in comparison with traditional teacher feedback, in addition to facilitating expressiveness and self-correction of errors. The researchers used intermediate-level English learners (the sample group included 60 students) who were randomly assigned to two groups: AI-generated feedback and human feedback provided throughout a six-week intervention period. Information was gathered by use of pre- and post-writing assignments, pre- and post-perception surveys among learners, and semi-structured interviews. Grammatical accuracy was checked in the writing samples. Quantitative data were reviewed to reveal significant differences between groups, and qualitative data were coded in the form of themes to examine the experiences and self-correction behavior of learners. Findings show that AI feedback had more beneficial effects on grammatical accuracy, whereas human feedback had more beneficial effects on reflective error correction and creative expression. The article points out the complementary role of AI tools and teacher instruction in the acquisition of holistic English writing proficiency.

Keyword: AI-generated feedback, human-generated feedback, Writing Accuracy, Creativity in Writing, Self-correction Behavior

Research Area: ELT

Country: Bangladesh

download