Generative AI in Higher Education: A PRISMA-ScR Scoping Review of Academic Support, Multilingual Mediation, Ethics, Dependence, and Risk

This PRISMA-ScR-informed scoping review mapped the emerging literature on generative artificial intelligence (GenAI) in higher education, with particular attention to academic support, multilingual mediation, ethics, dependence, risk, and implications for international students.Searches were conducted in Scopus, Web of Science, Google Scholar, KCI, and RISS for English-and Koreanlanguage sources published or available online from 2023 to March 2026.Records were screened through title/abstract and full-text review against predefined eligibility criteria, and 40 papers were retained from 65 initially identified records.Data were charted by source group, year, study design, population, country/context, primary theme, and explicit focus indicators.A structured analytical coding framework was then applied to map eight cross-cutting domains: student use/adoption, ethics/risk/integrity, multilingual mediation, measurement/scale development, international-student adaptation, AI literacy, institutional guidance, and dependence/reliance.The evidence base was concentrated in 2025, and international sources accounted for most of the included studies.Ethics, risk, and academic integrity constituted the largest macro-theme, while international student adaptation, multilingual support, institutional guidance, and dependence remained comparatively underdeveloped.The review advances existing GenAI-in-higher-education literature by integrating adoption, support, language mediation, ethics, dependence, and adaptation into a single evidence map.The findings support the need for context-specific empirical studies that examine GenAI not only as a productivity tool or misconduct risk, but also as a mediated support system for multilingual and internationally mobile students.