Abstract
This study examines how task-technology fit (TTF) and perceived usefulness (PU) are associated with responsible generative AI use (RGU) in higher education. Survey data from 280 university students in China were analyzed using covariance-based structural equation modeling. The model showed excellent fit, and all hypothesized paths were positive and significant: TTF → PU (β = 0.360), TTF → RGU (β = 0.318), and PU → RGU (β = 0.273). Bootstrap results indicated a partial mediation pattern in which PU was associated with the relationship between TTF and RGU. Females reported higher RGU than males, doctoral students outscored undergraduates on TTF and PU, and frequent GenAI users scored highest across all constructs. The findings extend TTF and TAM by integrating responsibility as a behavioral outcome and indicate that task-aligned, transparent GenAI practices may support more sustainable learning.