SHARP Retreat 2026: Towards joint theory building, Open Science principles, and Secure AI for research

One personal highlight was the Open Science kick-off chaired by Markus Gödker and Mario Gollwitzer featuring discussions on preregistration, registered reports, replication, reproducibility (games), and “switch to open” strategies for whole labs, a recent idea advanced by LMU’s Open Science Center.

On a conceptual level, Markus Gödker, Constanze Richters, and Doris Holzberger led an impressive joint theory-building activity involving more than 70 researchers. As a first step, we discussed and eventually refined the shared definitions of core SHARP constructs (e.g., representational scaffolding, learner model, adaptivity) in small groups around a "concept lead". A SHARP ontology will be the first milestone.

Building on several of these concepts, a workshop by Stefan Ufer introduced a joint design model (AdProSim) for describing and implementing adaptivity in simulations. Now the real work begins: applying AdProSim across contexts and studies. We also launched two new task forces: Transfer into medical and teacher education, led by Bernhard Schmidt-Hertha, and Science Communication of SHARP results to a broader public, led by Michael Sailer.

On the infrastructure side, Enkelejda Kasneci announced a secure AI server for SHARP members to run experiments on adaptive support for simulation-based learning, including GPT-OSS models. Good to know that we can fine-tune, prompt, and employ Large Language Models for experimentation and data analyses that run at Technische Universität München and keep our research data in safe places “intra muros” of the participating universities.

Speaking of data privacy and security: Stephan Hachinger introduced a strategy and infrastructure for SHARP’s research data management, supported by the Leibniz-Rechenzentrum.

Just before leaving the island, another remarkable event happened: led by Alexander Kacina, a foundational meeting for a new laboratory association took place. Twelve laboratories across the participating universities will share lab space, equipment for eye tracking, virtual reality, video and behavioral observation in teacher education, medical simulation centers, and DigiLLabs — and, of course, expertise within SHARP. The next step will be a shared booking system. We returned from the island somewhat tired but inspired.