A Validation of the Human-Generative Artificial Intelligence Trust Scale

Trust is a key concept in human–computer interaction, critical to the acceptance and use of generative artificial intelligence technologies. This study aimed to develop a trust scale tailored for human-generative AI interactions by adapting the original human–computer trust scale and evaluating its reliability and validity. An online survey was conducted with 310 valid responses. Results showed the scale consists of three dimensions—benevolence, competence, and reciprocity—comprising nine items, all exhibiting good discrimination. The scale’s overall Cronbach’s alpha was 0.891, with individual dimensions ranging from 0.806 to 0.810. Structural validity was confirmed (χ2/df = 2.32, TLI = 0.92, CFI = 0.95, RMSEA = 0.09, SRMR = 0.05), and criterion-related and convergent validity were well-supported. Based on the Human–Computer Trust Scale, the adapted Chinese version of the Human-Generative Artificial Intelligence Trust Scale exhibits strong reliability and validity, making it a valuable instrument for assessing trust in human-generative artificial intelligence interactions within a Chinese user context.