The Privacy Paradox and Algorithmic Resistance: Do Users Really Care About the Tailored Responses in Generative Artificial Intelligence

One growing trend in generative artificial intelligence (GenAI) consists of designing and providing tailored responses to help users access personalized services and drive users’ acceptance of algorithmic decisions. However, limited attention in prior research has been paid to the negative role of tailored responses on users’ algorithmic resistance in personalized interactions with GenAI. Based on the perspective of the privacy paradox and the privacy calculus theory, this study proposes an integrated model to investigate the impact of tailored responses on users’ algorithmic resistance in heterogeneous boundary conditions. One online experiment was performed to test the research model, and the results revealed and explored two distinct pathways: (1) the tailored responses of GenAI enhance the user’s perceived vulnerability, which in turn increases the possibility of users’ algorithmic resistance, and this pathway will be more evident in the context of judging tasks than in consultative ones; (2) the tailored responses can improve the user’s perceived relevance, which could affect the behavior of the algorithmic resistance, and a clear interaction impact between tailored responses and AI transparency with respect to perceived relevance was found. This study not only contributes to the literature on tailored services in GenAI but also provides practical insights for algorithmic designers and administrators of GenAIs.