Generative Artificial Intelligence in English as a Foreign Language (EFL) Higher Education: A Critical Review of Opportunities and Challenges

Abstract

Generative Artificial Intelligence (GenAI) is rapidly changing higher education, offering opportunities for English as a Foreign Language (EFL) teaching and learning while also posing pedagogical, ethical, and institutional challenges. This critical literature review examines recent research on GenAI implementation in EFL higher education. The review includes eighteen peer-reviewed publications, including empirical studies and systematic literature reviews, retrieved from Scopus, Google Scholar, and ScienceDirect between 2024 and 2026. The review indicates that GenAI has potential to support language learning through personalised feedback, adaptive learning, formative assessment, and increased learner engagement, with the most consistent evidence found for writing, reading, grammar, vocabulary development and formative feedback specifically. Evidence for other areas, including speaking and listening proficiency as well as benefits for learners at lower proficiency levels, is more limited and comes mainly from individual studies. The findings also suggest that AI-assisted learning can promote learner autonomy and self-regulated learning, and can support instructors in developing more student-centred instructional approaches. At the same time, the review identifies recurring challenges, including academic integrity, overreliance on AI-produced materials, data security, AI literacy, teacher preparation, and the need for clear institutional policy. These challenges appear to depend more on how GenAI is implemented than on the technology itself. The evidence suggests that GenAI is most effective as a tool that supports teachers rather than replacing them, provided it operates within appropriate pedagogical, ethical, and institutional frameworks. The paper concludes by discussing practical implications for EFL higher education, with particular reference to Benghazi University as an illustrative case, and by identifying priorities for further research, including longitudinal studies, broader geographical representation, and more direct investigation of learners at different English proficiency levels.