In generative AI we trust: an exploratory study on dimensionality and structure of user trust in ChatGPT

Abstract

The trust that users place in generative artificial intelligence (AI) can significantly influence their intentions and behaviors regarding its usage. Nonetheless, our current knowledge regarding user trust in generative AI is limited. To address this research gap, the study conducted semistructured interviews with 29 participants to investigate the factors that may influence user trust, using ChatGPT as an illustrative example. The findings led to the identification of several factors that account for user trust in ChatGPT. These factors encompass user-related aspects (such as technology attitude, innovativeness, and risk perception), information-related factors (including information source, information quality, and information values), technology-related factors (covering system quality, technology quality, and technology ethics), organization-related factors (encompassing organizational structure, brand reputation, and cultural context), and environment-related factors (involving policy environment, network environment, and social environment). In conclusion, the study formulated and presented the ChatGPT user trust framework as a theoretical model to comprehend user trust in generative AI. A thorough discussion of the framework is provided along with suggestions for future applications.