At a time characterized by an unprecedented flow of information, the European public is caught in a complex web of narratives, some of which drift dangerously into the realms of misinformation and disinformation. Combating disinformation remains challenging, but the targeted identification of misleading narratives offers hope. The project supports the work of the European Narratives Observatory by using innovative methods to combat misinformation.

Funded by the European Commission, the project will use so-called Large Language Models (LLM) to analyze large-scale text data and identify misleading narratives previously overlooked by older tools. The approach combines established techniques such as dynamic network analysis using LLM, which promises efficient application across different sources and languages. The focus is on key information hubs such as media and also Wikipedia, as these have a significant influence on public opinion. In addition, this study includes three in-depth case studies on the Ukraine conflict, LGBTQI+ narratives and the 2024 European elections, e.g. in France, the Baltic region, Italy and Romania. The researchers will produce 36 journalistic articles showcasing the techniques used and launch a MOOC for journalists.

The Erich Brost Institute is working on the study together with the other partners and will contribute the German part.