UPDF AI

A SYSTEMATIC LITERATURE REVIEW OF GENAI ASSISTED WRITTEN FEEDBACK IN SECOND LANGUAGE WRITING IN HIGHER EDUCATION

Ziyao Ji,Naginder Kaur

2026 · DOI: 10.35631/ijmoe.829068
International Journal of Modern Education · 0 Citations

TLDR

GenAI-assisted feedback in L2 writing is most effective in improving language accuracy, expanding vocabulary, and supporting revision, but its impact on more advanced writing skills, such as logical coherence and sentence structure, remains inconsistent.

Abstract

Generative artificial intelligence (GenAI) has advanced rapidly in recent years. Powerful systems such as ChatGPT have begun to reshape how written feedback is provided in second language (L2) writing courses at universities. This report presents a comprehensive analysis of 33 empirical studies published between 2022 and 2025. It examines the characteristics of GenAI-supported written feedback, its classroom applications, and its impact on L2 writing. The review follows the PRISMA guidelines and draws on studies indexed in Scopus, Web of Science, and ScienceDirect. It analyzes research trends, methods of use, participant backgrounds, feedback types, and reported learning outcomes. The results show that research in this area has grown rapidly since 2024. Most studies adopt mixed-methods approaches. They mainly focus on students learning English as a foreign language (EFL) in higher education institutions in Asia and the Middle East. GenAI-assisted feedback in L2 writing is most effective in improving language accuracy, expanding vocabulary, and supporting revision. However, its impact on more advanced writing skills, such as logical coherence and sentence structure, remains inconsistent. The research findings reveal that students generally respond positively to GenAI feedback. They value its quick responses, ease of use, and ability to provide personalized support. These features can also increase learners’ motivation and sense of autonomy. Some studies find that GenAI feedback can be as effective as feedback from teachers or peers in improving writing performance. However, it often lacks the detailed pedagogical guidance and ethical oversight provided by human instructors. Future research should place greater emphasis on long-term instructional designs and more effective prompting strategies. It should also examine more closely the cognitive and metacognitive processes involved when students revise their writing with the support of GenAI