Validity and reliability of the Chinese version of Threats of Artificial Intelligence Scale (TAI) in Chinese adults

Abstract

AbstractWith the outbreak of the COVID-19 pandemic, artificial intelligence (AI) has been widely used in fields such as medical treatment, while the threat of artificial intelligence has also received extensive attention. However, this topic has been only limitedly explored in China. To provide a measurement tool for AI threat research in China, this study aimed to examine the validity and reliability of the Threats of Artificial Intelligence Scale (TAI) in two Chinese samples of adults (N1 = 654, N2 = 1483). Results of exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) suggested that the one-factor model of TAI as the best fitting model. Furthermore, the Chinese TAI was significantly related to Positive and Negative Affect Scale and Self-Rating Anxiety Scale, proving good criterion-related validity of the Chinese TAI. In sum, this study suggested the Chinese version of the TAI as a reliable and effective tool in assessing AI threat in the Chinese context. Limitations and future directions are discussed.