Towards an XAI alignment workshop: a practice-oriented, multi-stakeholder approach for human-centred AI explanations

The field of Explainable AI (XAI) aims to find ways to create understandable and transparent AI systems to promote fairness, prevent bias and to make dealing with the systems easier. In our research we expand upon the existing approaches from a living labs and human-centred-design perspective. We present the XAI alignment workshop and its train-the-trainer concept as the first step of a holistic development methodology for context-specific explainable AI. Our aim is twofold: we focus (1) on how to create and embed AI explanations in a human context and (2) on how to support practitioners in doing so, since most systems that will affect the lives of people are created by professionals within a business context. To bridge the gap between research and practice, we are developing the workshop through case studies, and we created the train-the-trainer concept to try to empower practitioners to take ownership over the materials and build a more context-specific, inclusive and transparent way of working within their AI projects.