Turkish adaptation of the artificial intelligence ethics scale (EAI): a validity and reliability study for nursing...

This study was designed to culturally adapt the "Attitude towards Artificial Intelligence Ethics (EAI)" scale into Turkish and to evaluate its validity and reliability in the Turkish population.

This study was designed as methodological research to adapt the Attitudes Towards EAI Scale into Turkish and to evaluate its psychometric properties. The linguistic and cultural adaptation of the scale was carried out using the translation-back translation method. The study sample sample consisted of 656 undergraduate nursing students studying at a university in Türkiye. Participants were determined by a simple random sampling method, and voluntary participation was taken as a basis.

When the findings were evaluated, the results of the Exploratory Factor Analysis (EFA) showed that the original five-factor structure of the scale (transparency, harmlessness, privacy, responsibility, and fairness) was largely preserved. According to the Confirmatory Factor Analysis (CFA) results, it was determined that the five-factor model had good fit values. The scale's internal consistency was evaluated with Cronbach Alpha (α) coefficient, and α values for all sub-dimensions ranged between 0.85 and 0.95. These results show that the scale has a high level of reliability. In addition, composite reliability (CR) values above 0.70 and average variance extracted (AVE) values above 0.50 supported the convergent and discriminant validity of the scale. Furthermore, discriminant validity was further confirmed by MSV (< 0.31) and ASV (< 0.18), and EFA results via the FACTOR program (RMSEA = 0.047, CFI = 0.993, NNFI = 0.985) also supported the five-factor structure. Measurement invariance across gender was established at all levels (∆CFI < 0.01).

The results show that the EAI is an appropriate tool for assessing university students' attitudes toward artificial intelligence ethics in Turkish society.

In this direction, testing the scale in different professional groups, age groups, and cultural contexts in the future may expand its generalizability and usage areas. In addition, such scales are thought to provide important contributions to educational programs and policy development processes to increase ethical awareness in today's rapidly developing artificial intelligence technologies.