Self-reported AI literacy – more than just usage

Technological openness and emotional attitudes as predictors of self-reported AI literacy amongst students on social sciences and education degree programmes

Abstract

This study examines students’ self-perception of AI literacy and its relationship with technological readiness, emotional attitudes towards AI, and the frequency of use of AI tools amongst students on social sciences and education programmes in Upper Austria. The sample comprises 179 participants from three higher education institutions in Upper Austria (University of Applied Sciences Upper Austria [FH OÖ], Private Pedagogical University of the Diocese of Linz [PHDL], Johannes Kepler University [JKU]), who were surveyed online during the 2024/25 winter semester using the Meta AI Literacy Scale (MAILS) as well as supplementary scales on willingness to adopt technology and AI usage. The descriptive results show that 82.7% (n=156) of students use AI tools at least several times a week, indicating a high prevalence and integration of AI applications in everyday student life. In a hierarchical multiple regression comprising three models, technological readiness (β=.32, p<.001) and an enthusiastic attitude towards AI (β=.36, p<.001) emerged in the final model as significant positive predictors of self-reported AI literacy. A concerned attitude towards AI showed a negative, though statistically insignificant, association (β=−.13, p=.07). The frequency of use of AI tools (β=−.11, p=.85) and the students’ gender (β=.02, p=.82) were also not significant predictors. The overall model, which additionally included age as a control variable, explained 38 per cent of the variance in AI literacy (adj. R²=.38, F(6,124)=14.10, p<.001); Age proved to be a significant negative predictor (β=−.20, p=.006), meaning that younger students tended to report higher levels of AI literacy. The study highlights the importance of fostering an open attitude towards technology and high-quality engagement with AI during degree programmes in order to sustainably strengthen the AI literacy of future social and educational scientists. Finally, the study’s limitations and implications for higher education practice are discussed.