Trust Considerations for Explainable Robots: A Human Factors Perspective

Explanations alone don't guarantee appropriate trust: The paper emphasizes that simply making robots "explainable" doesn't automatically lead to healthy human-robot relationships. Trust must be carefully calibrated—users can over-trust or under-trust robots even when given explanations, which can harm team performance and safety.

Different bases of trust matter for different situations: The paper identifies that trust in robots can be built on different foundations (such as competence, reliability, or benevolence), and explainable robots should be designed to address the specific trust basis most relevant to their intended use, rather than assuming one explanation approach works universally.

Trust should be specific to actual robot capabilities: A key finding is that "trust specificity" is critical—users should trust robots only for the particular tasks and conditions they're actually designed for. Explanations should help users understand these boundaries rather than creating general confidence in the robot.

We need better ways to measure trust in explainable systems: The paper identifies that existing metrics for assessing whether explanations actually build appropriate trust are limited, suggesting the field needs standardized methods to evaluate whether explainable robots are truly improving human-robot team performance.

Human factors expertise is essential for XAI design: The paper argues that AI and robotics researchers should integrate insights from human factors and psychology literature when designing explanations, rather than focusing solely on technical accuracy or completeness of information.