Bridging the Human–AI Trust Gap: The Mechanisms Underlying Perceived Efficacy of AI-Powered Psychological Counseling Large Language Models

Abstract

Artificial intelligence (AI) is transforming the delivery of mental health services, yet public trust in AI-based psychological counseling remains a significant challenge. Drawing on a “technology–context–trust” framework, this study recruited a total of 1000 consumers in China through Credamo and conducted five between-subjects, scenario-based experiments to examine differences in perceived efficacy between LLM-based psychological counseling and human experts and to investigate the mechanisms underlying these differences. Perceived efficacy was the primary outcome of the study and, in Experiment 1B, also served as a mediator linking counseling provider type to downstream market outcomes, including overall attitude, willingness to try, and willingness to pay. The results showed that (1) consumers generally perceived AI as less efficacious than human experts, a difference driven by the parallel mediating effects of perceived professional competence and perceived risk; (2) participants rated the AI-based service as more efficacious in the routine counseling vignette, whereas they rated the human-expert service as more efficacious in the major psychological-disorder intervention vignette; (3) disclosure of technological investment enhanced trust in AI’s capabilities and mitigated negative evaluations; and (4) safety-standard disclosure reduced perceived risk, while higher perceived economic accessibility was associated with a smaller provider-related difference in perceived professional competence. These findings characterize subjective service evaluations among the recruited online consumers in China and do not establish comparative clinical effectiveness. They inform research on consumer trust while highlighting the need for further validation in more diverse samples and actual service settings.