The use of artificial intelligence in English language teaching and learning in higher education: a systematic literature review

This PRISMA-guided systematic review synthesizes empirical evidence on how artificial intelligence (AI) is used in English language teaching and learning (ELT/ELL) in higher education. A comprehensive search of Web of Science, Scopus, ScienceDirect and ERIC (2020–2025) identified 1,536 records; 101 peer-reviewed empirical studies met inclusion criteria and underwent mixed-methods synthesis combining thematic analysis, descriptive statistics and topic modeling. Results show a sharp increase in publications from 2020 to 2024, strong regional concentration and a marked imbalance across language skill domains. Mixed-methods designs prevail and samples are largely student-focused. AI tools cluster into five functional families and commonly act as teaching assistants, learning companions, and assessment/feedback providers. Reported outcomes are generally positive, but risks include over-reliance, variable feedback quality, limited cultural adaptability, and equity/ethics concerns. These findings indicate that effective AI integration in higher education ELT/ELL depends not only on tool adoption, but also on enhanced teacher preparation, culturally responsive AI design and longitudinal evaluation across diverse learning contexts.