Human-AI Interaction meets alpine research retreat

From January 17 to 23, 2026, members of the Human-AI Interaction group led by Prof. Sven Mayer at TU Dortmund University took part in the Winterschool 2026 at the Söllerhaus in Austria. The group was kindly invited by colleagues from LMU Munich, who have been organizing this Winter School series for more than ten years and continue to shape it as a central meeting point for the HCI and AI research community.
Set in a snow-covered mountain landscape with clear blue skies throughout the week, the Söllerhaus offered an inspiring and focused environment. Many participants started their days with skiing before coming together for intensive academic exchange. Being removed from everyday routines allowed everyone to concentrate fully on emerging research topics, free from the usual distractions of office life.

A strong presence from the Human-AI Interaction chair

This year, the newly formed Human-AI Interaction (HAII) team from TU Dortmund University attended the Winterschool as a group, including Ankur Bhatt, Lisa Townsend, Matthias Schmidmaier, Toan Nguyen, and future PhD student Fiona Lau. For the team, the Winterschool provided an important opportunity to connect with the broader European research community, establish new contacts, and further develop existing collaborations.
Together with around 50 participants - including PhD students, postdoctoral researchers, faculty members, and selected Master’s students - the Dortmund group became part of a diverse and highly motivated cohort. Participants came from a range of institutions, such as the University of Oslo, Lancaster University, and Humboldt University of Berlin. The mix of academic backgrounds and career stages encouraged open discussion and mutual learning.

Learning in an interactive environment

The program was structured around two thematic blocks. The first focused on applied AI in HCI research, addressing topics such as AI-supported writing, experimental design, reinforcement learning, and ethical aspects of AI-generated content. The second block explored multimodal generative models for text, images, and video, as well as applications in augmented reality, conversational agents, and wellbeing technologies.
A defining feature of the Winterschool was its highly interactive format. Many sessions emphasized group work, hands-on tutorials, and collaborative problem solving, enabling participants to actively engage with new tools and methods rather than merely observing them.

During the weekend, Prof. Albrecht Schmidt and Prof. Sven Mayer jointly led a hands-on session on Writing Papers with AI Support. Alongside practical insights into emerging tools, the session highlighted the continued importance of foundational academic skills - including the question of what actually makes a good abstract.
In addition, Prof. Sven Mayer contributed an introduction to reinforcement learning, with a particular focus on reinforcement learning from human feedback. This session helped participants gain a deeper understanding of key mechanisms underlying modern interactive AI systems and their relevance for human-centered machine learning.