Abstract

Generative artificial intelligence (GenAI) has diffused through higher education faster than the structures meant to govern it, yet acceptance research still treats adoption as a terminal outcome rather than as an event embedded in students' self-regulation. This study extends the Unified Theory of Acceptance and Use of Technology (UTAUT) by specifying metacognitive awareness (META) and self-regulated learning (SRL) as mediators between the four UTAUT determinants and behavioural intention (BI), and BI predicts self-reported use of GenAI as a cognitive partner (AU). Responses from 451 students at seven Thai Rajabhat universities were analysed with partial least squares structural equation modelling (PLS-SEM; SmartPLS 4; 10,000 bootstrap subsamples). Internal consistency was very high (above .95 for six of eight constructs), indicating partial item redundancy, and discriminant validity was supported under HTMT.90. Facilitating conditions, operationalised largely as perceived institutional guidance and support, showed the strongest associations with META (β = .595) and SRL (β = .514) but no significant direct association with intention (β = .061, p = .430); their only significant indirect association with intention ran through META (β = .186), a pattern consistent with indirect-only mediation. Social influence was the strongest direct predictor of intention (β = .357), followed by META (β = .312); effort expectancy and SRL did not predict intention. The model explained 62.1% to 70.9% of endogenous variance and showed out-of-sample predictive relevance. Given the cross-sectional, single-source design and full-collinearity diagnostics indicating possible common-method variance, the findings are interpreted as associations compatible with, but not establishing, the proposed mechanisms.

This work is licensed under a Creative Commons Attribution 4.0 License.