Keying Zhang, Sihui Li, Yating Huang
Generative artificial intelligence (GenAI) has introduced new opportunities for enhancing feedback practices in higher education. However, empirical evidence in this domain remains fragmented, providing limited insight into how GenAI-driven feedback shapes learning processes and experiences. Guided by the Student-Feedback Interaction Model, this scoping review analysed 51 empirical studies published between 2023 and 2025 to map current research on GenAI-driven feedback across four dimensions: feedback contexts and sources, messages, learners, and outcomes. The findings revealed that GenAI-driven feedback was investigated across varied educational settings, with increasing focus on human-GenAI integration and empirical analyses examining message characteristics generated by GenAI. The results demonstrated notable variation in students’ perceptions and engagement during GenAI-driven feedback processing, with evidence regarding both the immediate impacts and long-term developmental trajectories of such feedback remaining mixed. This scoping review identified critical directions for future inquiry: examining the effectiveness of human-GenAI integrated feedback across diverse contexts, strengthening the theoretical and pedagogical foundations of learner-centered prompt engineering, and exploring how student characteristics shape feedback processes and outcomes.