Trust and Acceptance in Human–Robot Interaction: Psychological Foundations for Industry 5.0

Abstract

This systematic review (PRISMA 2020) presents the results of 58 studies (2023–2025) on trust and acceptability in human–robot interaction (HRI) for Industry 5.0. Trust builds up slowly over the reliability of the robot, the robot's transparency, social appropriateness, and ethicality, but can collapse suddenly due to robot failure, ethical transgressions, or LLM hallucinations/inconsistencies, particularly for socially assistive robots and decision support robots. Over‐trust can result in complacency, automation bias, decreased monitoring, and safety risks, whereas under‐trust can result in non‐use, increased workload, and poor performance. Individualism–collectivism can also influence trust and trust recovery. Mitigation strategies for trust issues have been found to involve proactive disclosure, adaptive explainable AI (XAI), culturalized apologies, and behavioral trust recovery strategies. We ropose a human‐centered approach for strong and calibrated trust for Industry 5.0, which includes the integration of transparent XAI, human state monitoring, cultural adaptation, proactive interaction, and adaptive trust models.