DARPA formulated the Explainable Artificial Intelligence (XAI) program in 2015 with the goal to enable end users to better understand, trust, and effectively manage artificially intelligent systems. The four-year XAI research program began in 2017 and ended in 2021, and included multiple teams from academia and industry.
The program was organized into three major technical areas as illustrated below: (1) the development of new explainable models and explanation interfaces for generating effective explanations; (2) understanding the psychology of explanation by summarizing, extending and applying psychological models of explanation; and (3) evaluation of the new XAI techniques in two challenge problem areas: data analytics and autonomy.