In the past decade, the use of Generative AI as an e-learning assistant is gaining popularity among university students to support the completion of academic tasks. However, there is limited understanding of the factors that influence its continued use post-adoption. This study combines the Task-Technology Fit (TTF) framework and the Expectation-Confirmation Model (ECM) to evaluate the factors that influence continuance intention in the use of Generative AI by university students. A survey was conducted among 265 students from STEM and non-STEM fields of study, with data processing using the Partial Least Squares Structural Equation Modeling (PLS- SEM) method. The model explained 62% of continuance intention, with technology characteristics exerting the strongest influence on task-technology fit (β=0.638). Perceived usefulness (β=0.385) and satisfaction (β=0.387) emerged as the primary drivers of continuance intention, while moderation analysis showed no significant differences between STEM and non-STEM students. The results showed that perceived usefulness and satisfaction had a significant effect on the intention to continue using Generative AI. Technology characteristics also proved to play a greater role than task characteristics in influencing task-technology fit, which in turn significantly influenced the confirmation of usage expectations. However, moderation tests found no significant effect of field of study (STEM vs. non-STEM) on the relationship between perceived usefulness and intention to continue using Generative AI. These findings extend the theoretical understanding of the post-adoption behavior of Generative AI technology and provide practical implications for its application in the higher education sector. © 2025 IEEE.