Harnessing Generative AI for L2 Written Corrective Feedback: A Systematic Review of Current Research and Future Direction

Automated writing evaluation (AWE) has long been employed to assist second language (L2) learners in improving their writing through timely and consistent feedback. The emergence of generative artificial intelligence (GAI) introduces new possibilities for enhancing AWE systems. While a growing number of studies have explored using GAI for written corrective feedback (WCF), the findings remain mixed. This study systematically reviewed 37 empirical articles published between 2023 and 2025 that examine the application of GAI for L2 WCF. The review aimed to clarify research foci, examine implementation practices, and identify critical gaps to inform future research. Findings indicate that GAI tools show promise in providing accurate and helpful feedback, particularly at the surface level. However, limitations persist in addressing higherorder writing concerns. Hybrid feedback models that combine GAI with teacher input appear to yield the most comprehensive benefits. Additionally, issues related to student engagement, feedback uptake, prompt design, feedback sources, and individual learner differences require further investigation. This review highlights the need for more longitudinal, context-sensitive, and learner-centered research to fully harness the pedagogical potential of GAI in L2 writing instruction.