This study investigates the impact of Generative Artificial Intelligence (GAI) feedback types (directive vs. facilitative) and task types (structured vs. unstructured) on perceived usefulness and intention to use GAI among university students majoring in tourism. This study employs a 2 (feedback type: directive vs. facilitative) × 2 (task type: structured vs. unstructured) between-subjects experimental design. The results indicate that compared to directive feedback, facilitative feedback significantly enhances students’ perceived usefulness, which in turn positively influences their intention to use. Perceived usefulness fully mediates the relationship between feedback type and usage intention. Furthermore, task type moderates the effect of feedback type on perceived usefulness. Specifically, directive feedback is more effective in stimulating perceived usefulness under structured tasks, whereas facilitative feedback demonstrated greater advantage in unstructured tasks. This study contributes to the theoretical foundation of GAI applications in education by revealing how different feedback types and task characteristics influence university students’ willingness to adopt GAI, as well as the underlying mechanisms driving this relationship.