A Systematic Literature Review of Large Language Models in EFL

Abstract

As tech advances reshape education, large language models (LLMs) are becoming increasingly popular in language learning. Research shows LLMs like ChatGPT are widely used in education, mainly in higher education, with limited K-12 application. This paper examines their implementation and potential impact in K-12 education. Using a PRISMA-based systematic literature review, this study analyzed 2022–2025 empirical studies. It focuses on AI roles in LLM-supported second foreign language learning, covering educational domains, stages, and teacher roles. The findings offer insights for future research and practical suggestions for LLM integration in this field, advancing it and providing a reference for educators and researchers.