Calibrated vs. Overestimating Initial Information: Effects on Mental Model Accuracy and Performance in Simulated Human-Robot Collaboration

The present study investigated how initial information influences mental models and performance in human–robot collaboration. Participants ( N  = 61) completed a simulated search-and-rescue task after receiving either calibrated or overestimating descriptions of a robot’s capabilities. Initial information significantly affected both cognition and performance. Contrary to expectations, participants who received overestimating information showed smaller discrepancies in their collaborative mental models. However, these participants also demonstrated a greater decline in performance under higher task demands. No significant relationship was found between mental model accuracy and task performance. The findings suggest that initial information shapes human–robot collaboration through mechanisms beyond mental model accuracy alone. In particular, overestimating system descriptions may impair adaptation and collaboration efficiency as task demands increase. Our results highlight the importance of empirically evaluating onboarding and introductory system explanations when setting up human–robot collaboration.