Karl Jaspers and artificial neural nets: on the relation of explaining and understanding artificial intelligence in medicine

Abstract Assistive systems based on Artificial Intelligence (AI) are bound to reshape decision-making in all areas of society. One the most intricate challenges arising from their implementation high-stakes environments such as medicine concerns frequently unsatisfying levels explainability, especially guise so-called black-box problem: highly successful models deep learning seem be inherently opaque, resisting comprehensive explanations. This may explain why some scholars claim that research should focus rendering AI understandable, rather than explainable. Yet, there is a grave lack agreement concerning these terms much literature AI. We argue seminal distinction made by philosopher and physician Karl Jaspers between different types explaining understanding psychopathology can used promote greater conceptual clarity context Machine Learning (ML). Following Jaspers, we constitute multi-faceted epistemic approaches not seen mutually exclusive, but complementary ones themselves they necessarily limited. Drawing famous example Watson for Oncology highlight how Jaspers’ methodology translates case medical Classical considerations philosophy psychiatry therefore inform debate at centre current ethics, which turn crucial ethically legally sound medicine.