Revisiting TAM Under Institutional Diversity: Student Acceptance of AI-Based Learning Across Two Southeast Asian Contexts

Linda Indiyarti Putri (Faculty of Islamic Studies, Universitas Wahid Hasyim)
Sheerad Sahid (Faculty of Education, University Kebangsaan Malaysia)
Andy Prasetyo Wati (Faculty of Economics and Business, State University of Malang)
Ali Murfi (Faculty of Tarbiyah and Education, Sunan Kalijaga State Islamic University)

Jurnal Pendidikan Progresif
Vol 16, No 1 (2026): Jurnal Pendidikan Progresif

This study aimed to evaluate the measurement and structural models of the Technology Acceptance Model in the context of AI-based learning and to examine whether the hypothesized relationships vary across two university settings with different institutional characteristics. A quantitative survey design was employed involving students from two universities, namely Universiti Kebangsaan Malaysia (UKM) and Universitas Wahid Hasyim (Unwahas). Data were collected through a structured questionnaire covering the core constructs of the Technology Acceptance Model, including perceived ease of use, perceived usefulness, attitude toward use, behavioral intention to use, and actual system use. The data were analyzed using structural equation modeling and multi-group analysis. The analysis was conducted in two stages: first, the evaluation of the measurement model for reliability and validity; then, the assessment of structural relationships and cross-group variation. The measurement model demonstrated acceptable reliability and validity, indicating that the instrument was adequate for subsequent structural testing. In the overall structural model, all five hypothesized relationships were statistically significant, supporting the core sequence proposed by the Technology Acceptance Model. Perceived ease of use positively influenced perceived usefulness and attitude toward use; perceived usefulness positively influenced attitude toward use; attitude toward use positively influenced behavioral intention to use; and behavioral intention to use positively influenced actual system use. Multi-group analysis further revealed that the relationship between perceived ease of use and perceived usefulness was stable across both institutional groups. However, the pathways related to attitude formation and the translation of behavioral intention into actual use showed greater contextual variation. The findings confirm that the Technology Acceptance Model remains a relevant framework for explaining students’ adoption of AI-based learning technologies. At the same time, the results indicate that the strength of several acceptance pathways is partly shaped by institutional context. Therefore, AI implementation in higher education should be approached with sensitivity to differences in learning environments and institutional conditions. Keywords: artificial intelligence in education, technology acceptance model, measurement invariance, multi-group SEM, higher education.