Examining Generative AI User Intermittent Discontinuance from a C-A-C Perspective

Abstract

As a passive behavior, intermittent discontinuance may lead to user defection and undermine the continuous development of generative artificial intelligence. From a cognition-affect-conation perspective, this research examined the enablers and inhibitors of generative AI user intermittent discontinuance. We adopted a mixed method of structural equation modeling and fuzzy-set qualitative comparative analysis. The results indicated that privacy concern and information hallucination influence cognitive dissonance, which further leads to intermittent discontinuance. In contrast, perceived intelligence, anthropomorphism, and personalization influence affective commitment, which prevents intermittent discontinuance. The results imply that generative AI companies need to be concerned with both cognitive dissonance and affective commitment in order to prevent user intermittent discontinuance.