Social perception in Human-AI teams: Warmth and competence predict receptivity to AI teammates

Abstract

Advances in artificial intelligence (AI) promise a future where teams consist of people and intelligent machines, such as robots or virtual agents. In order for human-AI teams (HATs) to succeed, human team members will need to be receptive to their new AI counterparts. In this study, we draw on a tripartite model of human newcomer receptivity, which includes three components: reflection, knowledge utilization, and psychological acceptance. We hypothesize that two aspects of social perception—warmth and competence—are critical predictors of human receptivity to a new AI teammate. Study 1 uses a video vignette design in which participants imagine adding one of eight AI teammates to a referent team. Study 2 leverages a Wizard of Oz methodology in laboratory teams. In addition to testing the effects of perceived warmth and competence on receptivity components, Study 2 also explores the influence of receptivity components on perceived HAT viability. Though both studies find that perceived warmth and competence affect receptivity, we find competence is particularly important for knowledge utilization and psychological acceptance. Further, results of Study 2 show that psychological acceptance is positively related to perceived HAT viability. Implications for future research on social perception of AI teammates are discussed.