
Margarete Jahrmann is a ludic artist, researcher and professor Experimental Game Cultures at the University of Applied Arts Vienna. 2025 she received the Austrian media art price 2025, 2020 Media Art Prize of the City of Vienna. 2026 she showed at Chemnitz Kosmos Festival, e-commemorate Berlin. In 2025 she had a major solo-show at the Vienna AIL gallery. Further exhibitions: 2024 Kuandu Museum, Taipei and re:publica, 2023 AMAZE, 2021 invited artist at Biennale Habana. 2026 she published a monograph, KOPFGELD and Other Ludic experiments, neurosciences and art and THE TREE GAME at transcript. In her collaborative and interdisciplinary arts practice, she developed the LUDIC method, explores human and non- human, cognitive and the emotional linked to ecologies, democracy and society.
Co-Authors:
Thomas Brandstetter
Stefan Glasauer
ROBOPSY – An Artistic Examination of Role-play with LLMs in the Museum Context
FROG 2026 – Talk
The easy accessibility and wide availability of Large Language Models (LLMs) is fundamentally changing how people acquire information, especially about historical events. The distortions, limits and errors in the LLMs’ presentation of facts are well known in the academic discourse, but less so in the general public.
In recent times, historical museums have emerged as a vibrant platform for the discussion of questions dealing with the representation of knowledge: they have applied self-critique to the traditional ways of presenting their collections, have introduced new and interactive formats of museum communication and have established themselves as a forum for critical examination of the portrayal of historical events. All of this makes them ideal spaces for bringing a critical examination of LLMs to the public.
We propose that the play and workshop formats developed in our research project ROBOPSY would be a perfect fit for historical museums. ROBOPSY has created a game device that lets participants play a role-playing game with LLMs that is centered on a historical event. It allows a direct comparison between different LLM platforms, such as ChtGPT, Grok, Mitral or DeepSeek, and an easy customization and localization of the topic.
In our talk, we will present the installation, explain how it works and how it already has been used in different contexts. We will examine two different scenarios we have run, one about the murder of Moritz Schlick in 1936 at an exhibition at the AIL in Vienna and one on the so-called “Kirschenrummel” revolt in 1920 at the Atelierhaus Schaumbad in Graz. The latter was the result of a workshop session in which the participants could prompt the LLM and play the game together.
Our most recent experience took place in former Karl-Marx-Stadt about the Wende Prompt in Chemnitz Kosmos Festival, at a historic site work, at Kraftwerk Museum.
We will present our experiences with those formats and show how our device could serve as a ludic incubator for critically engaging museum visitor with the problems and pitfalls of LLMs. We also suggest that a historical museum might be the ideal place for the device, as the museum exhibits would provide a powerful context for a critical examination and comparison of different sources of information.
