AI4DH Lecture: Large Language Models for Analyses in Digital Humanities

On Monday, 26 January 2026, we launched the AI4DH 2026 Workshops and Lecture Series with a lecture by Prof. Dr Marko Robnik-Šikonja on large language models for analyses in digital humanities.

The lecture was intended for researchers in the humanities and social sciences as an introduction to large language models (LLMs): what they are, how they function, and how they can support research in digital humanities.

Why LLMs Matter for Digital Humanities

Large language models are reshaping methodological practices across many academic disciplines, including linguistics, social sciences, literary studies, historiography, journalism, political science, and folkloristics. Their ability to process and interpret large volumes of text makes them powerful tools for uncovering complex patterns and generating valuable insights from textual or visual data.

Understanding How LLMs Work

Dr Marko Robnik-Šikonja explained how LLMs work, focusing on their training on vast text and/or visual corpora to predict the next token. While this enables remarkable performance, it also means that LLMs are not immune to hallucinations and biases. The lecture emphasised the necessity of a human-in-the-loop approach, robust methodological design, and thorough evaluation to ensure trustworthy and meaningful outcomes.

Practical Adaptations: Fine-Tuning and Prompt Engineering

The lecture also examined two main methods by which LLMs are adapted for specific research tasks: fine-tuning, in which a pretrained model is further trained on domain-specific data and prompt engineering, in which carefully crafted prompts guide the model’s responses.

Evaluating Models for Complex Research Tasks

The lecture also highlighted the importance of establishing trust in model performance, particularly when analysing complex phenomena. The discussion outlined evaluation strategies that combine methodological rigour and transparency to equip researchers with tools to critically assess and validate their findings.

Coming Up: February 2026 Lectures and Workshops

Two additional lectures and workshops are scheduled for February 2026 as part of the AI4DH Lectures and Workshops Series. Find more information and register here: