From digital readiness to AI acceptance: A PLS-SEM study of mediation pathways in university administration

<ns3:p>Background Artificial intelligence (AI) integration in higher education requires understanding adoption factors among non-academic staff. This study examines how established technology acceptance models can assess AI readiness in universities, identifying key factors influencing AI use intention among administrative personnel. Methods An empirical survey of 213 non-academic staff at a Hungarian university employed validated constructs: digital readiness, AI readiness, perceived usefulness, perceived ease of use, facilitating conditions, and behavioural intention. Data were analysed using partial least squares structural equation modelling (PLS-SEM) examining direct and mediation effects. Results While general digital readiness was high, preparedness for complex AI applications was significantly lower. AI use intention was strongly influenced by facilitating conditions (β = 0.185, p < 0.05) and perceived usefulness (β = 0.602, p < 0.001). Digital readiness, AI readiness, and perceived ease of use showed no significant direct effects. A strong indirect pathway emerged from digital readiness through AI readiness and perceived usefulness to use intention (β = 0.071, p < 0.01). Conclusions Technology acceptance models can be adapted for AI assessment with modifications. AI adoption depends more on organizational support and perceived value than general digital readiness. Institutions must develop targeted AI competency initiatives beyond basic digital literacy programs.</ns3:p>