Trust is Contagious: Social Influences in Human-Human-AI Team

Abstract

When working in teams, individuals’ trust can be influenced by their teammates consciously or unconsciously through social interactions, a phenomenon defined in this paper as trust contagion. We investigated the effects of trust contagion in human-human-AI teams using a 2 (AI reliability: high vs. low, within-subjects factor) x 3 (confederate trusting: high, low, neutral, between-subjects factor) mixed-subject design. A team of three, consisting of one participant, one confederate, and one AI teammate, performed a ten-round trust-based game that requires resource allocation and joint decision making with their human and AI teammates. Results showed that when teaming up with high-trusting confederate, people showed higher trust in the AI teammate and performed better in the game. These findings suggest that social influences in human-human interactions can significantly affect human-AI trust, providing important theoretical implications for integrating AI in team settings.