AI adoption and AI governance in higher education: A BERTopic-assisted computational scoping review of risks, readiness, and policy responses
Generative artificial intelligence (GenAI) has intensified attention in higher education to institutional governance while concerns about individual adoption and readiness remain salient. This article maps the retrieved literature through a BERTopic-assisted computational scoping review of 1518 English-language journal-article records published between 2021 and 2026 and retrieved from OpenAlex. A Python/Google Colab workflow combined SentenceTransformer embeddings, UMAP dimensionality reduction, HDBSCAN clustering, and BERTopic class-based TF-IDF topic representation, yielding 14 substantive topics and 455 outlier records (30.0% of the corpus) under the selected metadata-augmented configuration. Adoption and perceived usefulness remain prominent, alongside institutional guidance, assessment redesign, academic integrity, AI literacy, and responsible GenAI use. A thematic synthesis of 34 accessible full-text papers informs a conceptual Higher Education AI Governance Readiness Framework integrating pedagogical readiness, AI literacy and capability, academic-integrity governance, assessment redesign, institutional policy clarity, and accountability and data governance. The study offers a corpus-level thematic map and an explicit boundary between computational and interpretive evidence; the proposed institution-level framework remains conceptual and has not been validated as a maturity, diagnostic, or causal model.
Authors
- Hoai Thu Le (ORCID: https://orcid.org/0009-0004-8500-3813)
- Hai Ninh Nguyen (ORCID: https://orcid.org/0000-0002-1658-531X)
Institutions
- Foreign Trade University (VN)
- Trường Đại học Khoa học và Công nghệ Hà Nội (VN)
- Hanoi University of Science and Technology (VN)
Publication Details
- Journal
- Social Sciences & Humanities Open
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1016/j.ssaho.2026.104015
- Primary Topic
- Artificial Intelligence in Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00