Have you ever wondered how artificial intelligence can be used to analyze historical sources? Similar to the natural sciences, the rapid progress of digitalization is also leading to fundamental changes in work processes in the humanities. This development goes beyond the mere storage and presentation of scientific data and is increasingly influencing the research process itself by transforming the methods of knowledge generation. In particular, the introduction of large language models (such as ChatGPT) has revolutionized historical research processes. Historical sources no longer need to be laboriously and time-consumingly analyzed manually; instead, the information can now be extracted automatically.
With our participation in the Long Night of Science, we would like to present the possible applications of artificial intelligence in the research of our interdisciplinary working group, consisting of computer scientists and historians. Our group is called MEPHIsto ("Digital Models, Explanations and Processes in the Historical Sciences"). We are investigating the administrative records of the late medieval church, which are published in the Repertorium Germanicum. This source offers a rich and multi-layered picture of the late medieval church, including its religious and canonical-legal foundations, organizational and fiscal mechanisms, personal and spatial networks as well as political ambitions and everyday realities.
As an example, we show how we used large language models to analyze payment transactions between Germany and the Roman Curia in the 15th century under Pope Sixtus IV. In this way, we were able to algorithmically identify a clearly structured network of important financial players of the time.
The hands-on offer gives you the opportunity to analyze historical sources yourself and find out whether your own ancestors were involved in the processes of the late medieval church. Use state-of-the-art Large Language Models to discover exciting connections. In addition, we offer a grammar construction kit mini-game in which you can parse a small German or Latin semi-structured text.