BIFOLD-HEIBRiDS Lecture

Natural Language Processing in times of ChatGPT

Abstract: The field of NLP has undergone fundamental changes since the introduction of Transformers in 2017 and the release of ChatGPT in 2022. Language models have evolved within a few years and show remarkable zero-shot capabilities, i.e., off-the-shelf models are able to reliably solve a large set of tasks without further fine-tuning. We will discuss these technological landmarks and what it means to be an NLP academic in 2025. The increase in model size makes it nearly impossible to train a large language model as a researcher in a public institution. At the same time, many models are being released by big corporations without details on training corpora or training procedures. Many models are even only accessible through APIs. Yet there is still a need for public NLP research and subfields like Explainability and Fairness became even more relevant. In this talk, I will give an overview of NLP and discuss some of the recent developments.

Short-Bio: Stephanie Brandl is a tenure-track Assistant Professor at the Copenhagen Center for Social Data Science (SODAS) at the University of Copenhagen. Her research focuses on fairness and explainability in natural language processing (NLP), as well as the application of language models to questions in the social sciences.