Can researchers in the field of humanities speed up and scale up their research by making use of AI-driven aides? Dr. Marianne Groep-Foncke of Rotterdam School of Management (Erasmus University Rotterdam) presented her research during the International Symposium on Grids and Clouds 2025. Her presentation was titled ‘Easily accessible LLMs in historical studies: opportunities, limitations, pitfalls‘.

In her talk, she explored if it is already possible to use simple and open-access Large Language Models (LLMs) to take over some of the more painstaking tasks, thereby leaving more time for the actual analysis of historical data. With her presentation, she took the audience, mainly consisting of experts in the field of computer science, on a journey along the historical Common Rules project of the research group Social Enterprises and Institutions for Collective Action. Sketching how the Phi-3-mini LLM performs in the categorization of the rules – a task that formerly was done manually – Groep-Foncke showed what the current generation of AI-driven models are capable of, but also highlighted the limitations of such models when working with historical data.