Medical Assistance Systems

Julian Leichert presented the paper “Influence of Prior and Task-Generated Emotions on XAI Explanation Retention and Understanding” at the MAI-XAI Workshop (2nd Workshop on Multimodal, Affective and Interactive eXplainable Artificial Intelligence) and attended the ECAI Conference in Bologna (28th European Conference on Artificial Intelligence, October 25–30, 2025).

The workshop focuses on exploring innovative methods for making data and models explainable while addressing the challenges of real-world applications, particularly in the context of multimodal affective interaction. It emphasizes not only the development of new XAI (Explainable Artificial Intelligence) approaches but also the creation of open resources, corpora, and benchmark datasets. These resources are essential for training models, evaluating their explanations in real decision-making processes, and ultimately developing powerful AI systems that are both human-centered and environmentally conscious.

The paper investigates how pre-existing emotions and task-induced arousal influence the understanding and retention of explanations provided by an explainable decision support system (DSS). The findings show that emotions such as anxiety can impair comprehension—particularly when they coincide with arousal generated by the task itself.

Contributors: Birte Richter, Christian Schütze, Anna Aksonova, Julian Leichert und Britta Wrede