Purpose This study examines generative artificial intelligence addiction (GAIA) among generative artificial intelligence (GAI) users to understand its causes, improve user experience and promote GAI sustainability. Design/methodology/approach Using the interaction of person–affect–cognition–execution (I-PACE) model, 563 valid responses were analyzed employing structural equation modeling (SEM) and fuzzy-set qualitative comparative analysis (fsQCA). Findings SEM results revealed that technophilia (TE), perceived task-technology fit (PT) and perceived inspiration (PI) enhance flow experience (FE), whereas TE, social isolation (SI), PT and PI lead to uncontrolled use of GAI (UUG). FE functions as a critical mediator for UUG, and both FE and UUG significantly contribute to GAIA. Additionally, fsQCA identified five configurational pathways to GAIA, highlighting its multifaceted nature. Originality/value This study represents one of the first attempts to apply I-PACE model to the GAIA phenomenon, thereby advancing the theoretical understanding of behavioral addiction among GAI users. By integrating SEM and fsQCA, this research highlights the pivotal role of UUG in GAIA development. To promote the responsible and sustainable growth of GAI, this study provides novel insights along with four actionable managerial recommendations.
Xu et al. (2026) studied this question.