X-RAI: A Framework for a Transparent, Explainable, Responsible, and Accurate Use of Artificial Intelligence

This article presents the X-RAI (an acronym for transparent (X-RAY), Explainable (X), Responsible (R), and Accurate (A) Artificial Intelligence (AI)) Framework. The purpose is to provide a framework enabling responsible management and documentation of AI systems throughout their entire lifecycle. The intended audience is professionals working with AI in an organizational context. The X-RAI Framework was designed, developed, used, and evaluated during a six-year-long collaboration between the Danish Business Authority and the IT University of Copenhagen.