When AI Delivers the Message: Emotional and Relational Consequences of AI-Generated Feedback in Marketing Education

Artificial intelligence (AI) is increasingly embedded in instructional feedback processes in marketing education, yet little is known about how students psychologically process feedback when it is attributed to AI rather than a human instructor. To clarify the state of the field, we first conduct a systematic literature review of 25 peer-reviewed studies on AI use in higher education, revealing a dominance of technology adoption and performance-focused models and a limited integration of emotional and relational mechanisms in AI-mediated feedback contexts. Building on source credibility theory and affective appraisal perspectives, we then test a moderated serial mediation model using a 2 × 2 experiment with undergraduate marketing students (N = 371). Results show that feedback valence significantly influences positive emotional responses, and this effect is amplified when feedback is attributed to AI rather than to a human instructor. Emotional responses subsequently shape perceived feedback clarity, instructor likability, and perceived instructor benevolence, which in turn predict engagement intentions. The findings demonstrate that AI-generated feedback is not processed as a neutral delivery channel but as a socially interpreted evaluative source. By integrating systematic synthesis with experimental evidence, this study advances a process-oriented understanding of the emotional and relational consequences of AI-mediated feedback in marketing education.