From Personal Concerns to Societal Threat Perceptions: Evaluating Generative AI Response in the Context of Risk

초록

This study examines how individual-level concerns about generative AI (GenAI) influence evaluations of GenAI responses through the mediating role of perceived societal threat. Specifically, AI-related competitive anxiety, job displacement risk, perceived hallucination risk, and privacy concerns are identified as key variables, and structural equation modeling is used to empirically assess how these individual-level perceptions extend to perceived societal threat. The results indicate that job displacement risk, perceived hallucination risk, and privacy concerns significantly increase perceived societal threat, whereas AI-related competitive anxiety does not show a significant effect. In addition, perceived societal threat positively predicts evaluations of GenAI responses. These findings suggest that users’ evaluations of GenAI operate across multiple levels and that macro-level societal risk perceptions and micro-level evaluations of AI-generated responses function together. In other words, users may perceive GenAI as socially threatening while simultaneously holding favorable evaluations of its responses. This study contributes to the literature by identifying a multi-level cognitive structure in which individual concerns extend to societal threat perceptions, as well as by demonstrating the presence of an ambivalent attitude structure. From a practical perspective, the findings highlight the importance of communication strategies aimed at reducing the gap between societal risk perceptions and user experience.