Artificial Intelligence Literacy and Responsible Use In Higher Education: A Structural Model Of Behavioral Intention Among Ecuadorian University Students
Artificial intelligence literacy is often assumed to be a uniform antecedent of its adoption, without empirically contrasting the relative weight of its distinct dimensions on behavioral intention to use. This study examined, through partial least squares structural equation modeling (PLS-SEM), the relationships between AI awareness, critical evaluation, ethical perception, self-efficacy, and prior experience, and behavioral intention to use artificial intelligence tools, in a sample of 300 Ecuadorian university students. The measurement model showed adequate reliability and validity across the six constructs assessed (AVE between 0.560 and 0.757; maximum HTMT of 0.634). The structural model explained 45.3 percent of the variance in behavioral intention, with eleven of the twelve hypothesized paths reaching statistical significance. The central finding of the study is that ethical perception toward artificial intelligence did not significantly predict behavioral intention to use it (beta = 0.042; p = 0.365), unlike awareness, critical evaluation, self-efficacy, and prior experience, all of which did. This result evidences a disconnect between ethical sensitization and declared technological behavior, with relevant implications for the design of training strategies on responsible artificial intelligence use in higher education.