Teacher vs. AI-generated feedback: students’ perceptions and impacts on learning

Purpose: Advances in generative artificial intelligence (GenAI) are reshaping feedback practices in language education, yet questions remain about how students meaningfully engage with AI-mediated feedback. While tools such as ChatGPT can offer immediate, interactive input, research highlights challenges, including limited student engagement, accuracy concerns, and the risk of overreliance. To address these, innovative learning environments that integrate diverse feedback sources with structured tasks and guided reflection are needed to support feedback literacy. In response, the present study implemented a feedback-focused instructional design in an English as a foreign language (EFL) teacher education course in Türkiye. The study examined how student teachers perceived and used teacher and AI-generated feedback.Design: Student teachers (n = 38) received both teacher and ChatGPT feedback and engaged in structured reflection activities in which they critically compared and responded to these feedback sources. Using a convergent mixed-methods design, data were collected through questionnaires, interviews, and reflective writing.Findings: Findings showed that participants perceived the two feedback sources as supporting their learning in distinct yet complementary ways. Participants valued ChatGPT feedback for its immediacy, clarity, and usefulness for initial revision, whereas they perceived teacher feedback as more trustworthy, contextualised, and pedagogically meaningful. Although participants appreciated the benefits of AI-generated feedback, they relied more heavily on teacher feedback when revising their work.Originality: The study suggests that combining ChatGPT feedback with guided reflection may support student teachers' feedback literacy by encouraging critical engagement with feedback. It also emphasises the potential importance of developing AI literacy and hybrid feedback practices in classroom-based assessment.