Artificial intelligence attitudes in higher education: Spanish validation of the AIAS-4 and its associations with well-being and academic AI usage

Abstract Amidst the swift expansion of artificial intelligence (AI) in education, which affects teaching, learning, and psychological engagement, there is a conspicuous absence of rigorously tested measures for assessing Spanish speakers’ attitudes toward AI. This study validates the Artificial Intelligence Attitude Scale (AIAS-4) in a sample of 650 university students from Spain (67.2% female; mean age = 22.8), exploring its psychometric properties and its ability to predict academic AI tool use. The scale was tested at two time points, three months apart, to assess reliability and longitudinal trends in AI tool usage. Confirmatory factor analyses, multigroup measurement invariance testing, t-tests, correlational analyses, and hierarchical regression models were conducted. Results indicated lower attitudes and usage habits among female participants. Core analyses confirmed the scale’s unidimensionality, reliability, gender measurement invariance, and convergent/discriminant validity, and revealed a modest but significant incremental prediction of academic AI tool use beyond common factors and covariates. Overall, the AIAS-4 appears to reliably assess students’ attitudes toward AI, to modestly predict their adoption of AI tools, and to help identify potential gender-gap shifts in educational contexts. These findings offer a validated tool to assess students’ evolving perceptions of AI, helping detect resistance early and guiding targeted interventions in education.