RCTrust Retreat 2026 in Berlin

After we had a great time at CHI 2026 in Barcelona, our research groups from RCTrust at TU Dortmund University, the University of Duisburg-Essen, and Ruhr University Bochum headed straight to Berlin for the annual RCTrust Retreat. The retreat focused on bringing our interdisciplinary research center together to shape our collective understanding of AI systems.

Lab Visits and Networking in Berlin

We did not waste any time getting started. On our first day, the group split up to explore some of Berlin’s most iconic research places: Charité Institute for Artificial Intelligence in Medicine, the Hasso Plattner Institute for Artificial Intelligence and Intelligent Systems, the Hasso Plattner Institute for Human-Computer Interaction, and the Max Planck Institute for Human Development.

The HAII group and the Inclusive Technology and Collective Engagement Group (led by Prof. Giulia Barbareschi) headed out to Potsdam to visit the Hasso Plattner Institute for Human-Computer Interaction, hosted by Prof. Dr. Patrick Baudisch. Prof. Baudisch and his students gave us a deep dive into their latest work, and seeing their progress with laser cutting and high-speed 3D printing in person was honestly eye-opening.

Prof. Mayer and Prof. Barbareschi also shared what we’ve been working on. It turned into a lively discussion, where everyone was just bouncing ideas off each other.

Main Event

The Retreat kicked off with a guest talk by Frederike Kaltheuner titled "Can you trust what you can't choose?" She shifted the conversation toward the infrastructure layer of AI, making a compelling case for why trustworthiness is not just about the interface we see, but the systems hidden underneath.

The following two days, the agenda gave PhD Students to have the opportunity to present their work in presentations and posters, on which they got plenty of feedback, as well as Barcamp sessions to explore new research ideas.

Aria shared her early PhD work on how different AI roles shape human behavior, especially AI as an intermediary in human-to-human communication and as an advisor using personal and contextual insights. I’m exploring how these roles affect authenticity, decision-making, and self-understanding. It led to some really interesting discussions about how AI influences the way we think and act.

Soumik, in his poster, tries to pinpoint the challenges of perceived authorship in Human + AI collaboration. Literature implies that creative credit becomes shared, ambiguous, and context-dependent, especially contextual information of the user's knowledge. Through a focus group study, this work plans to find user pain points in collaborating with AI.

Toan talked about Cognitive-aware neuroadaptive interfaces using EEG and eye-tracking, which have wide applications in education and healthcare, but the current measurement method of cognitive load does not capture their underlying sources. This project addresses that gap by separating these sources and using them to adapt specific interface components, enabling more precise and effective neuroadaptive systems.

Ankur presented his poster about his PhD topic titled “Human-Intent Prediction in Human-Robot Collaboration.” In his poster, he talked about Robots being implemented as Reactive and how we can make the Robots proactive that anticipate humans’ next action. He discussed four modules to predict human intent. First, human behavior understanding, second, activity labeling, third, predicting the intent using few-shot adaptation, and fourth, explainable robot action.

Barcamp Session and Brainstorming

After the formal presentations, we shifted gears into Barcamp sessions. Fiona, Prof. Dr. Nicole Krämer, and Prof. Dr. Markus Pauly led a session on the “Risk of Cognitive Deskilling”; the problem statement of Fiona's poster and Barcamp session addresses the risk that increasing the delegation of knowledge work to AI may lead to human cognitive deskilling. She discussed this issue with researchers from RCTrust, who also recognized its importance. These discussions sparked the idea of conducting a comprehensive RCTrust study, as well as organizing a Barcamp session. During the session, participants exchanged diverse perspectives on the topic and collaboratively brainstormed potential research directions and paper ideas.

Of course, it wasn't all data and delegates. We took some time on our second-to-last day to walk along the Berlin Wall, soaking in the history of the city.

Overall, the retreat was the perfect blend of high-level academic "nerding out" and genuine team bonding. We’re heading back to our respective universities with full notebooks, new research partners, and a lot of momentum for the rest of 2026.

Until next time, Berlin!!