Analysing Explanation-Related Interactions in Collaborative Perception-Cognition-Communication-Action

Humans primarily want clarification about decisions and actions: The study found that most explanation-related messages in collaborative tasks seek clarification about why teammates made specific decisions or took particular actions, rather than requesting other types of explanations. This suggests AI systems should prioritize being able to justify their choices clearly.

Explanations directly affect task performance: The research confirms that explanation-related communications have a measurable impact on how well teams complete their work, indicating that robots without adequate explanation capabilities could undermine collaborative effectiveness, not just user satisfaction.

AI robots need explanation abilities to earn trust: The paper emphasizes that AI-equipped robots working alongside humans must be able to explain their behavior to cooperate effectively and build trust—suggesting that explainability isn't just a nice-to-have feature but fundamental to human-robot teamwork.

Emergency response contexts reveal explanation needs: By studying a simulated emergency response task, the research identifies real-world scenarios where explanation capabilities matter most, providing practical guidance for where to prioritize AI explainability development in high-stakes collaborative settings.

Explanation types should match human expectations: The study bridges explainable AI literature with actual human communication patterns, helping developers understand which types of explanations humans naturally expect from teammates—and therefore which capabilities should be built into collaborative AI systems.