The influence of agent and message type on perceptions of social support in human-machine communication

Abstract

In this study, we investigated the differences in perceived competence, credibility, and social presence within socially supportive interactions (emotional or instrumental) using human, AI, and social robot actors. We further differentiated our findings by actor and type of support and found that participants thought the human and social robot differed in terms of competence but did not rate them differently on impressions of support. Our findings suggest that people may be more attuned to the outcomes of support versus the actors involved in the interaction. In this brief report, we present CASA as a theoretical framework, offer a detailed analysis of our findings, and discuss the perceived criteria for effective socially supportive interactions.