Unpacking the rejection of L2 students toward ChatGPT-generated feedback : an explanatory research

Abstract

Purpose: Drawing on the Unified Theory of Acceptance and Use of Technology (UTAUT), this sequential mixed-method study explored L2 learners’ challenges in seeking corrective feedback in an AI-assisted educational environment. Design/Approach/Methods: The data was collected from 45 university students in a Computer Science program who were encouraged to seek corrective feedback from ChatGPT for their argumentative writing report. Self-reflection data were qualitatively and quantitatively analyzed. Findings: Findings suggested that 45.9% of the AI-generated feedback was rejected by students, with a higher rejection rate observed for content-focused feedback (58.7%) compared to form-focused feedback (41.3%). AI-generated content-focused feedback was rejected due to misalignments between expected and received feedback, anticipated high workload to respond, mismatch between feedback from AI and external references, and impeding conditions for students to engage with the feedback. Meanwhile, the rejection of form-focused feedback was largely associated with the high level of expected effort. Originality/Value: This study is one of the few existing studies on the L2 learners’ challenges in seeking corrective feedback from generative AI systems. Notably, the study differentiates between form-focused and content-focused feedback and identifies the distinct challenges and learning opportunities that impact students’ rejection of AI feedback.