Validacija skale stavova prema ChatGPT-u
Validation of the Attitudes Scale towards ChatGPT

Summary/Abstract: The rapid development of large language models such as ChatGPT has significantly impacted various aspects of everyday life and industry. Given the growing importance of ChatGPT, there is a need to examine attitudes toward this type of technology. The aim of this study is to develop and validate the ChatGPT Attitude Scale. Two studies were conducted with a total sample of 491 participants. The first study involved 160 participants (Mage=28.47, SD=10.09), and the second study included 331 participants (Mage=27.50, SD=8.70). To verify the factor structure of the scale, an exploratory factor analysis was conducted in the first study, identifying a three-factor structure. The total explained variance was 66.29%, with the components accounting for 43.80%, 15.14%, and 7.35% of the Usage, Concern, and Usefulness variance, respectively. The internal consistency coefficients were α=.89 for the first, α=.84 for the second, and α=.85 for the third component. Convergent validity was confirmed using the Technology Readiness Index and the Personal Innovativeness in Information Technology Scale.The rapid development of large language models such as ChatGPT has significantly impacted various aspects of everyday life and industry. Given the growing importance of ChatGPT, there is a need to examine attitudes toward this type of technology. The aim of this study is to devel¬op and validate the ChatGPT Attitude Scale. Two studies were conducted with a total sample of 491 participants. The first study involved 160 participants (Mage=28.47, SD=10.09), and the second study included 331 participants (Mage=27.50, SD=8.70). To verify the factor structure of the scale, an exploratory factor analysis was conducted in the first study, identifying a three-fac¬tor structure. The total explained variance was 66.29%, with the components accounting for 43.80%, 15.14%, and 7.35% of the Usage, Concern, and Usefulness variance, respectively. The internal consistency coefficients were a=.89 for the first, a=.84 for the second, and a=.85 for the third component. Convergent validity was confirmed using the Technology Readiness Index and the Personal Innovativeness in Information Technology Scale. Personal innovativeness in infor¬mation technology positively correlated with Usage (r=.60, p<.01) and Usefulness (r=.38, p<.01), and negatively with Concern (r=-.29, p<.01). Regarding the Technology Readiness Index, Usage positively correlated with Innovativeness (r=.49, p<.01) and Optimism (r=.45, p<.01), and nega¬tively with Discomfort (r=-.20, p<.01) and Insecurity (r=-.32, p<.01). Similarly, Usefulness posi¬tively correlated with Innovativeness (r=.40, p<.01) and Optimism (r=.38, p<.01), and negatively with Discomfort (r=-.24, p<.01) and Insecurity (r=-.30, p<.01). Concern negatively correlated with Innovativeness (r=-.27, p<.01) and Optimism (r=-.24, p<.01), and positively with Discom¬fort (r=.33, p<.01) and Insecurity (r=.53, p<.01). In the second study, confirmatory factor analysis confirmed the proposed three-factor model of the scale. The chi-square correction relative to the degrees of freedom (x2/df = 2.55) indicated an acceptable model fit. Other fit indices also supported the model's acceptability: GFI (.92), RMSEA (.07), CFI (.95), TLI (.94), SRMR (.05), and NFI (.92). The results demonstrated that the ChatGPT Attitude Scale has a stable three-factor structure and satisfactory psychometric properties. Understanding attitudes toward various forms of artificial intelligence is crucial for the successful adoption and integration of new tech¬nologies into society.

• Details
• Contents