This training session is an adapted version of past courses on AI frailties and explainable AI. I set the focus lens on legal aspects of explainable AI, and the current gap that exists within regulatory bodies and actual practices. The course is comprised of a two hour-long theoretical background with a two hour-long lab session. The first part of the lab session focuses on post-hoc xAI methods and aim to outline their limitations. The second part of the lab session introduces case-base reasoning models.