What is Missing in XAI So Far?

Abstract With the perspective on applications of AI-technology, especially data intensive deep learning approaches, need for methods to control and understand such models has been recognized gave rise a new research domain labeled explainable artificial intelligence (XAI). In this overview paper we give an interim appraisal what achieved so far where there are still gaps in research. We take interdisciplinary identify challenges XAI point open questions with respect quality explanations regarding faithfulness consistency explanations. On other hand see interaction between user allow adaptability specific information needs explanatory dialog informed decision making as well possibility correct by interaction. This endeavor requires integrated rigorous approaches empirical evaluation based psychological, linguistic even sociological theories.