Blaming yourself, your partner, or an unexpected event: Attribution biases and trust in a physical coordination task

Abstract

As robots enabled by artificial intelligence become more agentic, people may come to develop trust schemas based on a robot's actions and attribute blame to the robot as they would with a human partner. Trust and blame have yet to be investigated during dynamic physical coordination tasks despite the potential ramifications for manufacturing and service industries that could benefit from effective human–robot physical coordination. In anticipation of future human–robot work configurations, we developed a joint physical coordination task as a preliminary test environment for understanding trust and blame in a work partner. Fifty-five participants were asked to jointly balance and transport a weighted box along a fixed path, and we used this test environment to evaluate the impact of a surprising event on trust in a work partner, and attribution of blame following a negative performance outcome. Results indicate that the group who experienced a surprising event compared to the group who did not trusted their partner more, but there was no difference in the attribution of blame to themselves, their partner, or to the surprising event. Conversely, the group who did not experience a surprising event tended to blame themselves for the negative outcome. These findings suggest that environmental uncertainty may prompt people's attribution of blame across multiple parties, including themselves. Moreover, people may build trust in work partners through the shared experience of surprising events. Future work would benefit from adopting our study design to investigate whether these findings are extendable to human–robot joint actors.