Measurement invariance of the artificial intelligence attitude scale (AIAS-4): cross-cultural studies in Poland, the USA, and the UK

This study investigates the measurement invariance of the Artificial Intelligence Attitude Scale (AIAS-4) across three countries: Poland (N = 511), the UK (N = 230), and the USA (N = 300). The AIAS-4 is a concise measure of general attitude towards artificial intelligence developed by Simone Grassini. The validation process consisted of two steps. Firstly, the cultural adaptation of the Polish version of the AIAS-4 was conducted, which was necessary for subsequent cross-cultural comparisons. The factor structure, internal, and construct validity of the Polish AIAS-4 were estimated. Next, multigroup confirmatory factor analysis was conducted to compare measurement invariance across three populations: Polish, British, and American. The results confirm that the factor structure of the Polish version of the scale is identical to the original one, with high reliability and acceptable validity. Multigroup confirmatory factor analysis supports configural invariance and partial metric invariance. However, equivalence at the scalar level has not been achieved in cross-cultural research.