Cross-Cultural Variations in Trusting Generative AI: The Roles of Uncertainty Avoidance, Digital Literacy and Perceived Control

Abstract

Trust is a key element of generative artificial intelligence acceptance and responsible use, which could be different for different cultural settings and users. This study investigates the differences between the trust placed in generative AI across cultures by exploring uncertainty avoidance and digital literacy and perceived control. To understand how these factors affect users' trust in AI-generated content among different cultural groups, a quantitative comparative analysis was conducted. The analysis took into account differences in trust and technological ability, as well as differences in toleration for ambiguity and the extent to which the user felt they could verify, modify or reject the product of the AI. Their results suggest that uncertainty avoidance is linked to a reduced level of trust; when users are less tolerant of ambiguity, they are more guarded against the unpredictable or potentially incorrect responses from AI. There is a positive correlation between digital literacy and trust as greater knowledge of technology helps people to better comprehend the system's functions and limitations. Perceived control also boosts trust by enabling people to have control over the choices they make and to think about the recommendations made by AI. The study also reveals that trust should be adjusted and not maximised as it can have negative consequences in either overreliance or underreliance. It states that a culturally responsive design, clear explanations, user control and cultivating a digital-literacy are crucial for creating informed and responsible interactions with generative AI.