Human Centered AI

Friday 6 Feb 2026, 14:15 → 15:45 Europe/Berlin

DO: JvF25/3-303 | BN: b-it/1.047

Natalia Andrienko, "Human-Centred AI: Connecting Human and Computer Intelligence through Visual Analytics"

Human-Centred AI (HCAI) is a multifaceted field dedicated to ensuring AI systems amplify rather than replace human abilities. While definitions of HCAI vary across the research landscape, this session focuses on the Visual Analytics (VA) perspective: the science of analytical reasoning supported by interactive visual interfaces. VA can serve as a powerful enabler for HCAI by facilitating synergistic collaboration between human reasoning and machine computation and supporting model building, auditing, and explainability. We shall move on to a practical example demonstrating the role of VA in the model building phase, which includes data structuring, feature engineering, and iterative model refinement.

Bahavathy Kathirgamanathan, "Beyond Model-Centric AI – How do we get more human-centered AI explanations"

Traditional explainable AI (XAI) methods are generally model-centric where they seek to describe how a model arrives at its predictions, but often neglects whether those explanations are insightful, trustworthy, or usable for humans. In this presentation, we introduce some key concepts in Explainable AI and look into the importance of using interactivity to allow human-centric exploration of model explanations.