How mindsets shape trust in AI: A dual path mechanism of self-efficacy and risk perception

Abstract

With the rapid advancement of artificial intelligence (AI) technologies, their applications in daily life have become increasingly pervasive. In human-robot interaction, trust is considered a critical predictor for the adoption of new technologies. Drawing on the lens of risk, this study integrates two cognitive pathways—internal self-assessment and external risk evaluation—within the framework of protection motivation theory(PMT). A parallel mediation model is proposed to examine how a growth mindset influences human-robot trust through the dual mechanisms of risk perception and self-efficacy. In Study 1, data were collected through a survey, revealing that a growth mindset positively predicts human-robot trust, with self-efficacy and risk perception serving as parallel mediators. To further clarify the causal relationships, Study 2 employed an experimental design, manipulating participants' growth mindset and assessing their trust behavior in a trust game paradigm to validate the model. The results demonstrated that different mindsets led to significant variations in human-robot trust investment, reinforcing the parallel mediating roles of risk perception and self-efficacy. The findings elucidate the underlying mechanisms through which a growth mindset affects human-robot trust, extend the application of PMT to the domain of human-robot interaction, and provide new insights into the trust formation process in such interactions.