| Time: | November 20, 2025, 10:00 a.m. (CET) |
|---|---|
| Universitätstraße 34, Room 150 | |
| Download as iCal: |
|
We are pleased to announce our upcoming ELLIS Unit Stuttgart Distinguished Lecture Series talk by Matthias Kraus (University of Augsburg)! Looking forward to seeing you all there! No registration necessary.
Title: Towards Meta-Interactive AI for Human-Centred Proactive and Explainable Agents
Abstract: Artificial intelligence is increasingly capable of taking initiative by anticipating user needs, offering timely support, and even acting autonomously. Yet, when should systems act on their own, and how can such initiative strengthen rather than undermine human cooperation and agency? The first half of this talk explores the concept of human-centred proactivity: AI behavior that anticipates and assists while balancing trust and task effectiveness. Drawing on insights from human-computer interaction, cognitive psychology, and recent experimental studies, I will
examine how proactive systems influence trust, coordination, and shared decision-making. I will argue that true cooperation arises not from automation, but from mutual understanding and adaptive coordination between
humans and machines. This perspective sets the stage for the second half of the talk, which introduces Meta-interactive AI, a framework for explainability as an ongoing, participatory dialogue between users and autonomous agents.
Bio: Dr.-Ing. Matthias Kraus is a postdoctoral researcher at the Lab for Human-Centered AI, University of Augsburg. His work advances trustworthy, proactive dialogue systems and human–AI/robot collaboration. In 2024 he served as Interim Professor of HCI at LMU Munich. He received his Dr.-Ing. in a co-tutelle between Ulm University and the University of Granada, Spain, following M.Sc. and B.Sc. degrees in Electrical Engineering and Information Technology at Ulm University. Kraus has led grants on socially adaptive HRI (FORSocialRobots) and contributed to EU/DFG/BMBF projects. His work has been published at leading conferences on user modeling, multimodal interaction, human–robot interaction, and NLP.