BACKGROUND

In the past few years, the rapid development of artificial intelligence (AI) and its success in many areas of everyday life have attracted global attention. Some discussions have noted that generative AI tools can make decisions on their own with the potential to improve themselves. Subsequently, conspiracy theories have emerged concerning the future implications of AI. In the present study, the Artificial Intelligence Conspiracy Beliefs Scale (AICBS) was developed to assess conspiracy beliefs concerning AI, andits psychometric properties were examined.

METHODS

A cross-sectional survey was conducted with 788 Turkish participants (M= 25.10 years, 56% female). The sample was split to carry out an exploratory factor analysis (EFA; n = 423) and a confirmatory factor analysis (CFA; n = 365), resulting in a 30-item scale comprising five subdimensions.

RESULTS

The five-factor structure explained 62.58% of the total variance. The CFA showed acceptable model fit indices and confirmed the EFA's five-factor structure. Based on the EFA's factor loadings, a short five-item version of the AICBS (AICBS-5) was developed with one item from each subdimension (which explained 45.28% of the variance). The CFA confirmed the unidimensional structure of the AICBS-5. The internal consistency coefficients of the AICBS, its subdimensions, and the AICBS-5 demonstrated very good reliability. Correlation analyses with external criterion measures (AI Anxiety Scale, Generic Conspiracist Beliefs Scale-5, and Anomie) supported the concurrent validity of the AICBS, its subdimensions, and the AICBS-5.

CONCLUSION

The findings demonstrate that both AICBS and AICBS-5 are valid and reliable psychometric instruments to assess AI conspiracy beliefs.