Standardized Soul: A Mixed-Methods Study on the Efficacy and User Perception of AI-Augmented Peer Support (AAPS) in College Mental Health Support

Abstract

College students are facing a structural mental health crisis, particularly in non-WEIRD contexts where scarce campus mental health resources and cultural norms hinder help-seeking behavior. We propose AI-Augmented Peer Support (AAPS), a collaborative workflow where human volunteers refine AI-generated empathetic responses. Through a between-groups experiment (N = 122) and semi-structured interviews (N = 30), we compare AAPS with professional counseling, pure AI, and peer support. Quantitative results show AAPS achieves comparable emotional relief to professionals while demonstrating significantly greater performance stability. Qualitative research highlights how AI's characteristics, such as neutrality, alleviate evaluative anxiety and authority pressure in educational settings. These findings integrate a synergistic approach into tiered care system design, demonstrating how AI can empower trained volunteers as specialized tools. Ultimately, this work illustrates how human-AI collaboration can balance algorithmic efficiency with human warmth to provide scalable, culturally sensitive support, fostering a more empathetic and intelligent society.