Motivational and appraisal factors shaping generative AI use and intention in Austrian higher education students and teachers

Franziska Kinskofer^*Maria Tulis

• Paris Lodron Universitat Salzburg, Salzburg, Austria

This study replicates and extends the Unified theory of acceptance and use of technology (UTAUT) to examine factors influencing generative AI (genAI) use among Austrian higher education students (n = 3,094) and teachers (n = 1,767). Confirmatory structural equation modelling (SEM) replicated previous evidence including performance expectancy, effort expectancy and social influence. Exploratory partial least squares SEM (PLS-SEM) was introduced for challenge and threat appraisals as predictors. We assessed the validity of the model in Austria, including higher education teachers—a group underrepresented in prior research. Behavioral intention predicted genAI use (β = .75, p < .001 students; β = .48, p < .001 teachers), with performance expectancy, effort expectancy, and social influence as key positive predictors. Effort expectancy was particularly salient for teachers, reflecting their limited time. Gender differences emerged primarily in students’ motivational antecedents. Females reported lower subjective competence, intrinsic motivation, and challenge appraisals but higher threat appraisals in line with previous evidence concerning digital and AI competence; these differences were weaker in teachers. To explore the new predictors, we conducted linear regression analyses investigating effects of genAI-related subjective competence, trust in genAI, and intrinsic motivation on challenge and threat appraisals for students and teachers. Challenge appraisals, in the regression predicted by intrinsic motivation, trust in genAI, and genAI-related subjective competence, positively influenced behavioral intention, while threat appraisals—linked to low competence and motivation—had a small negative impact (β = -.03, p = .004 students; β = -.03, p = .014 teachers), highlighting the affective dimension of AI adoption. The extended model explained substantial variance in behavioral intention (R² ≈ .8 in both groups) and usage behavior (students R² = .34; teachers R² = .18). Our findings emphasize the importance of aligning AI facilitation with user needs, intrinsic motivation, and addressing affective responses to promote meaningful and ethical adaptive genAI integration within the often cognitively and affectively demanding and high-pressure environments of higher education. Future research should consider individual differences, institutional culture, and evolving AI landscapes (e.g., learning analytics) in correlational as well as experimental designs to optimize adaptive AI use among diverse educational stakeholders.