Using LLMs to update human rights datasets
Dec 1, 2025
·
1 min read
Image credit: Point Normal/Unsplash.
New artificial intelligence tools are being released at a staggering pace, making traditional ways of data collection look outdated and potentially obsolete. However, the real benefits and limitations of large language models in the context of human rights data collection are not yet very well understood. With this project, I am testing the use of LLMs for the extraction of incident data on human rights violations from UN reports, exploring opportunities and trade-offs of expanding datasets that were previously human-coded.
Authors
Researcher in Political Science
I am passionate about exploring and comparing human rights protection and state coercion in democratic as well as authoritarian contexts. For my work and studies, I have received various scholarships and awards, and spent considerable time abroad in countries on five continents. I was previously a visiting scholar at Stanford University, USA, and a research fellow at the Global Public Policy Institute, Germany, where I co-developed the Academic Freedom Index. I hold a Ph.D. in political science from the University of Freiburg.