Architecting the AI-Driven University: Empirical Evidence on Adoption Pathways from Malaysia and Indonesia
TL;DR
Findings indicate that satisfaction-focused AI strategy must prioritise ease-of-use improvements and workflow integration over training volume, while Indonesian institutions specifically need structural investment in infrastructure and governance frameworks as preconditions for effective capacity-building.
Abstract
Grounded in the Technology Acceptance Model (TAM), this study investigates what determines AI adoption behaviour and user satisfaction among staff and students in Malaysian and Indonesian universities. Survey data were gathered from 748 respondents across 12 institutions in 2025. Using binary logistic regression to identify predictors of training participation and k-means clustering to reveal latent user segments, we find that perceived usefulness and perceived ease of use are stronger predictors of both AI usage and satisfaction than formal training attendance. Malaysian respondents were 2.47 times more likely to have completed AI training than their Indonesian counterparts (p < .001), yet this structural advantage produced no significant difference in satisfaction between the two countries (Mann-Whitney U, p = .214). Three user profiles emerged: AI Skeptics (23.8%), who require demonstration of practical task value before any training engagement; AI Learners (42.9%), who benefit most from discipline-embedded mentoring and competency recognition; and AI Champions (33.3%), best deployed as peer facilitators rather than additional training recipients. For policymakers, these findings indicate that satisfaction-focused AI strategy must prioritise ease-of-use improvements and workflow integration over training volume, while Indonesian institutions specifically need structural investment in infrastructure and governance frameworks as preconditions for effective capacity-building.