Learning effectiveness in AI-supported higher education: examining the roles of AI feedback quality, human-AI collaboration, self-Directed learning, and digital readiness

The rapid advancement of generative artificial intelligence (AI) has transformed higher education by reshaping learning processes, instructional practices, and human-AI interactions. Grounded in Social Cognitive Theory, this study examines associations among AI feedback quality, human-AI collaboration, self-directed learning, learning performance, learning effectiveness, and digital readiness. A cross-sectional survey was conducted with 164 undergraduate students who had prior experience using generative AI tools in university learning activities, and the proposed associations were examined using structural equation modeling. AI feedback quality, human-AI collaboration, and self-directed learning were positively associated with learning performance, which was positively associated with learning effectiveness. Statistically significant indirect associations through learning performance and interaction effects involving digital readiness were also observed. These findings should be interpreted cautiously because learning performance and learning effectiveness were assessed with highly overlapping self-report items concerning the perceived benefits of AI-supported learning. The direct and indirect associations may therefore partly reflect semantic redundancy and shared method variance rather than distinct educational processes. Moreover, the cross-sectional, single-source design does not establish temporal mediation or causal effects. Subject to these limitations, the findings suggest that technological support, learner agency, and digital readiness are jointly relevant to students’ perceptions of AI-supported learning.