Psychometric Evaluation of the Turkish Version of the Perceived Moral Agency Scale for Artificial Intelligence

This study examined the psychometric properties of the Turkish version of the Perceived Moral Agency Scale for Artificial Intelligence. Two studies were conducted: the first involved exploratory factor analysis, parallel analysis, validity and reliability analyses; the second involved confirmatory factor analysis, test–retest reliability, and gender-based measurement invariance analyses. Findings confirmed a two-factor structure: morality and dependency. Regarding convergent validity, morality correlated positively with anthropomorphism and mind-attribution, while dependency correlated negatively with these variables. For criterion validity, morality showed a positive correlation with conspiracy beliefs. Internal consistency and test–retest coefficients indicated acceptable reliability for both subscales. The scale demonstrated measurement invariance (configural, metric, scalar, and strict) across genders. The scale may contribute to research on the relationship between perceived moral agency and variables such as trust, acceptance, and ethical perceptions, in education, where artificial intelligence is increasingly integrated into contexts such as AI tutors and AI-supported learning environments.