To be credible or to be creative? Understanding the mechanisms driving user satisfaction with AI-generated content from a dual fit perspective

Abstract

Generative artificial intelligence (GAI) can fundamentally disrupt how content is produced and is increasingly integrated into organizational and individual task-performing and decision-making. This study investigates how individuals perceive and process AI-generated content by advancing the understanding of user satisfaction through an extension of dual fit theory. By theoretically grounding our investigation in dual fit theory, we propose that two key information representations (information credibility and creativity) are key drivers of user satisfaction, operating through two distinct mediating mechanisms: cognitive fit, which reflects the alignment of information with task demands, and emotional fit, which captures the match between the interaction experience and user emotions. Our model posits that both credibility and creativity distinctly influence these cognitive and emotional fit pathways. Furthermore, we examine the boundary conditions by exploring the moderating role of task representation (e.g., task routineness and task hedonism). We tested our hypotheses through two 2 × 2 between-subject scenario experiments with 548 and 197 participants. Across both studies, the results consistently show that information credibility positively impacts both cognitive and emotional fit. Information creativity primarily enhances emotional fit, with its effect on cognitive fit emerging under specific task conditions. Both fit dimensions, in turn, positively affect user satisfaction with the outcome and the process. Our results indicate that task routineness is a stronger moderator than task hedonism. A good match between the information and task representations (i.e., aligning credibility with routine tasks and creativity with creative tasks) leads to enhanced cognitive and emotional fit.