A core assumption of explainable ai systems is that explanations change what
users know, thereby enabling them to act within their complex socio-technical
environments. Despite the centrality of action, explanations are often
organized and evaluated based on technical aspects. Prior wo