Published in Higher Education Studies in 2026, this scientometric study maps 824 Scopus-indexed AI–IoT education papers from 2016–2025. Annual output grew from one document in 2016 to 274 in 2024. The research landscape was fragmented and technology-led, with pedagogical terms peripheral to its strongest intellectual clusters. A proposed saturation signal was sensitive to incomplete 2025 indexing and should not be read as proof that the field has matured or stopped expanding.
Key findings
- The study mapped 824 Scopus-indexed records published between 2016 and 2025.
- Annual output rose from one in 2016 to 274 in 2024.
- Conference papers represented 47.1% of research documents in the study's classification.
- AI, IoT and machine learning dominated central clusters, with pedagogical concepts more peripheral.
- Apparent saturation depended on incomplete 2025 indexing.
Why this matters globally
The mapping can help identify research gaps and cross-disciplinary collaboration opportunities internationally, but is not a causal assessment of AI–IoT learning effectiveness in real classrooms.
Thai researcher contribution
The bibliographic record attributes authors to North Eastern University, Rajamangala University of Technology Isan and Rajabhat Maha Sarakham University. Their contribution is to international evidence mapping, not a direct intervention in Thai classrooms.
Limitations to consider
Coverage depends on Scopus indexing and the chosen search expression. Incomplete 2025 records distort time-series modeling. Publication counts and keyword clusters do not measure research quality or student learning outcomes.