Last update:
   11-Nov-2026
 

Arch Hellen Med, 44(1), January-February 2027, 78-82

ORIGINAL PAPER

Attitudes Towards Artificial Intelligence Scale
Translation and validation in Greek

O. Konstantakopoulou,1 A. Katsiroumpa,1 I. Moisoglou,2 P. Gallos,3 O. Galani,3 M. Tsiachri,1 P. Lialiou,4 K. Lamprakopoulou,5 P. Galanis1
1Laboratory of Clinical Epidemiology, Faculty of Nursing, National and Kapodistrian University of Athens, Athens
2Faculty of Nursing, University of Thessaly, Larissa
3Faculty of Nursing, University of West Attica, Athens
4Laboratory of Computational Biomedicine Research, Department of Digital Systems, University of Pireus, Pireus
5Laboratory of Embedded Systems Design and Applications, Department of Electrical and Computer Engineering, University of Peloponnese, Patras, Greece

OBJECTIVE To translate and validate the Attitudes Towards Artificial Intelligence Scale (ATAI) in a sample of the general population in Greece.

METHOD We employed the forward-backward method to translate and adapt the ATAI in Greek language. We examined the reliability of the ATAI by calculating Cronbach's alpha. Also, we performed a test-retest study to examine the reliability of the ATAI by calculating the intraclass correlation coefficient (ICC). We examined the construct validity of the ATAI by performing confirmatory factor analysis (CFA). We examined the concurrent validity of the ATAI using the Artificial Intelligence Attitude Scale (AIAS-4) and the Short Trust in Automation Scale (S-TIAS).

RESULTS We found that the ATAI had very good reliability since ICC for the factor "acceptance" during the test-retest was 0.985 (95% confidence interval [CI]: 0.968 to 0.993, p<0.001). Also, the ICC for the factor "fear" during the test-retest was 0.971 (95% CI: 0.940 to 0.986, p<0.001). Moreover, Cronbach's coefficient alpha for the factor "acceptance" was 0.797, and for the factor "fear" was 0.705. We found that the Greek version of the ATAI had the two-factor structure as the original version. All indices indicated an acceptable two-factor model. In particular, x2/df was 0.961, RMSEA was <0.001, GFI was 0.986, AGFI was 0.947, TLI was 1.000, IFI was 1.000, NFI was 0.972, and CFI was 1.000. Concurrent validity of the Greek version of the ATAI was very good since we found statistically significant correlation between the ATAI and the AIAS-4 (acceptance scale: r=0.762, p<0.001; fear scale: r=-0.151, p=0.113), and the S-TIAS (acceptance scale: r=0.648, p<0.001; fear scale: r=-0.140, p=0.140).

CONCLUSIONS The Greek version of the ATAI is a reliable and valid tool to measure attitudes towards artificial intelligence in the general population.

Key words: Attitudes Towards Artificial, Intelligence Scale, Artificial intelligence, Greek, Reliability, Validity.

© Archives of Hellenic Medicine