Explanation in human-agent teamwork

There are several applications in which humans and agents jointly perform a task. If the task involves interdependence among the team members, coordination is required to achieve good team performance. This paper discusses the role of explanation in coordination in human-agent teams. Explanations about agent behavior for humans can improve coordination in human-agent teams for two reasons. First, with more knowledge about an agent's actions and plans, humans can more easily adapt their own behavior to that of the agent. Second, with more insight in the reasons behind an agent's behavior, humans will have more trust in the agents, and therefore more easily coordinate their actions. The paper also presents a study in the BW4T testbed that examines the effects of agents explaining their behavior on human-agent team performance. The results of this study show that explanations about agent behavior do not always lead to better team performance, but they do impact the user experience in a positive way.