Person-Centered Support From Chatbots Versus Humans

Abstract: This study aims to examine whether and how perceived identity of a support provider and message quality would affect perceived agency of the provider and support outcomes. A total of 196 college students in China participated in a laboratory experiment. Each participant was instructed to discuss a personal problem with either a chatbot or a human who would provide higher or lower person-centered support as preprogrammed. The results showed that perceived identity of a support provider and person-centeredness of support messages jointly affected participants’ perceptions of the provider agency, which in turn positively influenced participants’ emotional distress reduction, reappraisal, and perceived validation. The findings provide theoretical and pragmatic implications for the application of human–computer interaction in supportive communication.