Trust as Extended Control: Active Inference and User Feedback During Human-Robot Collaboration - Salomon Lab