LLM for Philosophical Counseling - Study 3

This study examines how users evaluate philosophical counseling experiences when the counselor is presented as either a human or an AI system. Specifically, it investigates whether users’ judgments are shaped by the actual source of the counseling response (human-generated vs. AI-generated), the disclosed identity label of the counselor (human vs. AI), or the interaction between the two. Using a 2 × 2 between-subjects experimental design, participants are randomly assigned to one of four conditions and read a simulated counseling dialogue adapted from prior materials. After exposure, participants report their satisfaction with the counseling experience, perceived authenticity, and perceived professionalism or effectiveness. The study further examines potential psychological mechanisms underlying these evaluations, including perceived empathy, mind perception, shame related to help-seeking, and willingness to engage in self-disclosure, as well as individual differences in perceived social support and AI literacy. The goal of this research is to clarify how different presentations of human and AI counselors influence users’ perceptions and evaluations of counseling interactions.