Skip to content


Abstract of the Talk at the interdisciplinary FRIAS Lunch Lecture:

The question for trustworthy AI involves a number of philosophical topics. The presentation concentrates on the issues of explainability and interpretability of machine learning. Explainability and interpretability are commonly considered synonyms, defined in terms of each other, or divided along superficial differences. The presentation, in contrast, draws a clear distinction between the two often confused terms. This is important to dissolve common but misled formulations of problems of interpretability and explainability that result from the attempt to interpret processes that can only be explained. Yet, the presentation shows that not only explanation but also interpretation of machine learning is possible, although in a way different from what is usually attempted. Not only the AI system itself but also the environment into which it is embedded needs to be considered. Trustworthy AI needs to be not only explainable, but also interpretable in the newly understood sense. 

The event takes place via Zoom. All current information about the event series and participation will be posted at