“Let me try again” – Leveraging Social Interfaces to Improve Trust in Automation for Teleoperation

We explore if social strategies can be used more effectively by social or traditional interfaces to repair trust in shared autonomy robot teleoperation. Shared autonomy systems help robots and operators collaborate, but when trust in these systems is broken, it does not recover easily. Social human-robot interaction suggests that social robots can rebuild trust through social strategies. This study explores if these strategies can be also used by teleoperated robots, and further compares if these social trust repair strategies are more effective for an anthropomorphized interface (more similar to social robots), or a traditional log and read-out style interface design when used with a non-social, teleoperated robot. We find that positive social perception of an interface can improve the effectiveness of social trust repair strategies, and that while social interfaces can repair trust better than traditional interfaces, poor social perception of that social interface may lead to worse trust overall. Our results suggest social interfaces for teleoperation robots may act as a way to leverage findings in social robotics in teleoperated robots, hold implications for social interfaces, and calls for additional investigation into appropriate social behavior and design across software agents, social robots, and teleoperated robots.