Does AI Help EFL Learners? A Scoping Review of Effectiveness, Engagement, and Ethics Across 131 Scopus-Indexed Studies

Abstract

- This scoping review maps 131 Scopus-indexed publications (2022 – 2026) examining AI-mediated English language teaching (ELT). Following PRISMA 2020 guidelines (Page et al., 2021), we searched Scopus and analyzed retrieved records through bibliometric analysis, thematic synthesis, and keyword co-occurrence mapping. Publication volume grew from one article in 2022 to 70 in 2025, a trajectory driven by the emergence of generative AI tools such as ChatGPT. Thematic analysis identified four clusters: (1) effectiveness of AI interventions across vocabulary, writing, and speaking skills; (2) engagement outcomes encompassing motivation, anxiety reduction, and foreign language enjoyment; (3) the diversity of AI tools spanning chatbots, intelligent tutoring systems, and large language models; and (4) ethical concerns including academic integrity, AI dependency, and algorithmic bias. Geographic analysis revealed a pronounced concentration in Asia-Pacific contexts, with China accounting for 60 publications (45.8%). Given the breadth of topics and the heterogeneity of study designs across the corpus, this review adopts a scoping rather than systematic synthesis approach, mapping the landscape of current research rather than producing definitive effect estimates. The evidence base suggests promising but context-dependent effectiveness for discrete language skills and predominantly positive engagement outcomes; however, most primary studies were conducted in 2022 – 2024 and should be read as reflecting AI capabilities of that period rather than the current generation of tools. Significant gaps remain in longitudinal research, theory-driven investigation, and empirical examination of ethical issues.