The mechanisms of AI feedback’s influence on college students’ learning engagement in human-AI collaborative learning

This study explores how AI-generated feedback affects student learning engagement in human-AI collaborative contexts. A grounded theory approach was used to analyse data from interviews with 50 undergraduates. The resulting model ‘Adaptive Engagement Cycle with AI Feedback’ explains how students’ reasons for using AI influence their interpretation of feedback, which then influence their cognitive, emotional, and behavioural engagement. The process is also influenced by contextual factors, including students’ views on the limitations of AI and their comparisons between AI-generated and human feedback. The findings indicate that students process feedback actively rather than respond passively, and that the influence of AI feedback depends on a key stage in which students interpret the feedback and experience certain emotional reactions. These interpretations and reactions then influence how they continue to engage with their learning tasks. This perspective shifts the focus from the feedback itself to the student’s active role within human-AI collaborative learning.