There has been growing interest in EFL writing classrooms regarding the use of ChatGPT for giving feedback. Peer-delivered feedback, long celebrated for its collaborative and motivational advantages, now coexists with ChatGPT’s promise of instant, technically precise corrections. This mixed-methods study, designed as a convergent parallel design, tracked 48 Turkish university EFL learners across a semester, analyzing eight writing tasks reviewed by both peers and ChatGPT, alongside student reflections from structured written interviews. The results revealed that peer feedback excelled in nurturing coherence, structure, and idea development, while ChatGPT-generated feedback proved stronger for grammar and vocabulary corrections. Participants appreciated ChatGPT’s speed and objectivity but perceived it as tone-deaf to cultural nuance and rhetorical purpose. Although peer-delivered feedback was regarded as occasionally inconsistent, it provided authentic engagement and a sense of audience. Building on these converging findings, this study offers a reasoned pedagogical implication, rather than an empirically tested model, termed feedback sequencing : using ChatGPT first for technical cleanup, then peers for higher-order concerns such as clarity and impact. This staged proposal, which warrants direct experimental verification in future research, is theoretically grounded in cognitive load theory and sociocognitive perspectives on formative feedback. The findings urge educators to weave AI literacy into curricula, not as a replacement for peer interaction, but as a complement that preserves the irreplaceable human dimensions of learning to write.
Harun ÇİFTCİ (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: