Innovation in Language Learning

Written Corrective Feedback (WCF) plays a significant role in addressing errors and enhancing learners' linguistic abilities, and it remains a crucial component in second language (L2) writing instruction (Balla et al., 2025; Hongxia & Razali, 2025). However, delivering detailed and personalized feedback can be labor-intensive and resource-demanding, especially in large classroom settings (Mulenga & Shilongo, 2025). Recent advancements in artificial intelligence, particularly the advent of large language models (LLMs) like ChatGPT, offer promising avenues for automating and supplementing traditional feedback mechanisms (Gao et al., 2023). Prior research indicates that AI-mediated feedback supports learners’ linguistic development by offering timely corrections that complement teacher input. Hybrid approaches combining human and AI feedback have shown promise in improving outcomes and motivation. However, studies exploring learners’ and teachers’ perceptions of AI-driven feedback remain limited, with concerns about trust, pedagogical value, and dehumanization of learning. This study investigates the efficacy of hybrid feedback approaches that combine teacher input with ChatGPT-generated feedback, comparing it to teacher-only feedback. Forty adult EFL learners are randomly assigned to two groups: an experimental group (n=20) receiving hybrid feedback and a control group (n=20) receiving only teacher feedback. The study also explores learner and teacher perceptions regarding ChatGPT-mediated feedback, focusing on trust, satisfaction, pedagogical value, and attitudes toward integrating AI tools in language learning. Data collection includes pre-test, post-test, and delayed post-test writing assessments, plus questionnaires to assess both linguistic gains and attitudinal shifts. This research contributes to strategies for integrating AI into language education, enhancing learner engagement and scalable feedback mechanisms.