The potential for AI applications exists in all disciplines. "Without the use of AI, we work with random samples in history," says Kara Kuebart, a research associate at the Institute of History at the University of Bonn. "The data volumes are simply so huge that it's impossible to process everything manually. Only with machine learning methods can we fully evaluate the data sets."
To ensure that future historians are already imparted with AI know-how during their studies, Kara Kuebart has been developing seminars and exercises in this area since 2022. "The goal is to give them digital skills to create maps, digitally analyze texts, or process data using computer support," she explains. "I want to break down students' inhibitions about code-driven AI programs."
"Initially, I acquired a lot of knowledge on my own, starting with experimenting with the Python programming language," she explains. Then she learned about the "AI for All" course offered by the University of Bonn as part of the BnTrAInee project – a project that connects existing AI expertise in computer science with various disciplines to develop needs-based teaching and learning offerings.
The hard road to AI expertis.
"In the intensive courses, we teach participants the basics of artificial intelligence methods," explains Dr. Moritz Wolter, who designed the courses together with his colleague Dr. Elena Trunz. "The only requirement is that you have completed the Python tutorial beforehand."
It's generally a bit easier for natural scientists, "they're already familiar with the mathematical foundations from their studies." But humanities scholar Kara Kuebart has also persevered – and discovered a passion for machine learning. In the intensive courses, participants not only learn the mathematical foundations but also solve various programming tasks. "One of the most popular exercises is when participants train a language model using only Shakespeare's works," says Wolter. "The texts the model produces then sound like Shakespeare—only they're not as artistically demanding."
"Moritz and Elena have structured the courses very well. They go through the different types of AI, explain how to set them up, and provide an overview of what's possible," says Kuebart. "This helps you understand which AI you can use for which question."
She now passes on the knowledge she has learned to her students. "Interest is very high. In a recent course, we created maps using a programming library. Everyone had to write code for it. It worked really well." Kara Kuebart also uses artificial intelligence methods in her own research. "At our chair, we use AI, for example, to search through old editions of the Kölnische Zeitung from the 19th and 20th centuries, to filter out changes in mood in political articles about the First World War, or to analyze advertisements for maid positions: What were the discourses like back then? What qualities were sought after?"
So far, almost 100 researchers have participated in the courses. The "AI for All" offerings are constantly being developed further. "We are guided by the needs of our participants," says Wolter. "Over time, we will offer further advanced courses on more in-depth topics." Kuebart is certain: "If there are further advanced courses, I will participate again."
About the University of Bonn's Digital Strategy
The University of Bonn's Digital Strategy (www.digital.uni-bonn.de) defines the measures and structures for its digital transformation. The BNTrAinee project is a measure in the strategy's Digital Competencies target area and aims to develop AI competencies across disciplines. More about the digital strategy at https://www.digital.uni-bonn.de/de .