Generative AI in English Language Teaching: A Systematic Review of Its Impact on Teaching Practices and Learning Outcomes

Abstract

Generative artificial intelligence (GenAI) tools are now common in English classrooms, but the research on them is spread thinly across different tools, settings and skills, and most of it records what people think rather than what learners actually gained. This review looks at how GenAI is changing the way English is taught and what learners gain from it. Using the PRISMA procedure, the review drew on Web of Science, Scopus, ScienceDirect and Springer, selecting peer-reviewed journal papers from 2024 to 2026; 28 studies were analysed in full. In most studies GenAI served as a helper for teachers in lesson planning, materials, feedback and classroom activities, but how far it was taken up varied with teachers’ own attitudes, self-assurance and preparation. Assessment was the area most challenged, prompting teachers to rethink task design and academic integrity. In terms of learning, GenAI is associated with perceived improvements in writing, grammar, vocabulary, pronunciation, motivation, confidence, and autonomy, but most evidence is perception-based rather than measured. The review concludes that GenAI is best used as a supplement to teacher-guided instruction, with teachers checking its output and guiding learners towards critical use. Future research should adopt longitudinal and controlled designs that measure language outcomes directly.