Akademik Veri Yönetim Sistemi
Current Psychology, cilt.45, sa.5, 2026 (SSCI, Scopus)
With the rapid development of artificial intelligence (AI), AI anxiety is becoming a psychological construct that is associated with a number of other variables and is frequently reported in a variety of populations. This study aims to examine the reliability of the original version by Wang and Wang and the adapted version by Akkaya et al. of Artificial Intelligence Anxiety Scale, a prominent measure of AI anxiety. Meta-analytic reliability generalization of the original version of the AIAS was performed using Cronbach's alpha coefficients from 15 primary studies with a total sample size of 4975. For the the adapted AIAS version by Akkaya et al., the total sample size was 3407 and the number of primary studies was 10. Categorical moderator analysis of the obtained overall Cronbach's alpha coefficients according to Likert type, research type, scale language, study group, country, and study field variables was performed. Also, continuous moderator analysis was conducted for the ratio of women variable. When the overall alpha coefficients were examined, it was seen that all pooled Cronbach’s alpha coefficients were high. The pooled alpha coefficient for the overall scale of the original version of the AIAS is 0.949 [%95 CI = 0.944—0.963], while this value is 0.878 [%95 CI = 0.844—0.935] for the adapted AIAS version by Akkaya et al. Even the pooled alpha coefficients obtained in different variable categories were generally high. For the original 21-item AIAS, when the moderator analysis results are examined, it is concluded that the difference between the overall alpha coefficients in the country and study field variables is significant. In addition, it was determined that the ratio of women variable significantly predicted the alpha coefficient for the Job Replacement subscale. Finally, suggestions for future research and practice are presented along with relevant implications.