Do Explainers Perceive the Explainees’ Specific Interests and Knowledge?

In adaptive explanations, XAI should consider the users, their interests and their knowledge. In order to enable XAI to be capable of this, it is still necessary to learn more about human mental representations, which Explainers have of Explainees, and implement findings accordingly. According to the philosophy of technology, XAI is a technological artifact and therefore possesses a specific structure, comprisingof an objective Architecture and a subjective Relevance side: How is the XAI structured vs. what is the purpose of the XAI? The question is whether human Explainers can perceive the Explainees’ interests inand knowledge about a technological artifact in regard to this dual nature theory. Therefore, an experimental explanation study (N=65) in which technological artifacts were explained was conducted. Explainers encountered and explained to confederates, who acted as Explainees. Two conditions were set: The confederates explicitly showed interest in either Architecture or Relevance. A retrospective video recall method with six measuring points was used to allow Explainers to relive certain moments.Quantitative questionnaires were administered to assess their perception of the Explainees’ interests and knowledge. The Explainers’ perception of the Explainees’ interests differed significantly between conditions, partly also between measuring points. The Explainers’ perception of the Explainees’knowledge differed significantly between conditions and measuring points. We found that if one duality side is demanded by the Explainee the Explainer can perceive this. This finding is especially importantfor XAI as the dual nature is not only a theoretical construct but can be perceived and measured.Keywords: Mental Representations · User Model · Technological Artifacts · Human-Human Explanations · Experimental Study