Psychometric properties and Turkish adaptation of the artificial intelligence attitude scale (AIAS-4): evidence for construct validity

Abstract

Artificial intelligence (AI) attitude scales can be used to better evaluate the benefit and drawback cons of AI. This article consists of two different studies examining attitudes towards AI. In Study I (N = 370), the four-item Artificial Intelligence Attitude Scale-4 (AIAS-4) has a one-dimensional structure as a result of confirmatory factor analysis and the fit index values are at an acceptable level [Comparative Fit Index (CFI) = 0.991; Goodness of Fit Index (GFI) = 0.989; Normed Fit Index (NFI) = 0.988; Tucker-Lewis Index (TLI) = 0.973; Standardized Root Mean Square Residual (SRMR) = 0.018]. Additionally, according to the results of the item response analysis conducted to support construct validity at this stage, the scale items have sufficient discrimination (discrimination value range = 2.22-3.80). Later, measurement invariance analysis revealed that the scale measured the same construct in females and males. In Study II (N = 331), the reliability of AIAS-4 was reached by calculating different reliability coefficients. Then, AI attitude was found to be associated with depression, anxiety, and stress, as well as mental health variables such as mental wellbeing and flourishing. Moreover, openness to experience, conscientiousness, extraversion, and neuroticism are significantly related to an AI attitude. Lastly, psychological distress has a significant mediating role in the relationship between AI attitude and mental health. The findings of this pioneering research on AI attitudes were discussed and interpreted in light of the literature.