As malicious campaigns multiply, the PROMPT (Predictive Research On Misinformation & Propagation Trajectories) pilot project advances the work of the European Narratives Observatory. We analyze vast amounts of textual data to detect malevolent narratives, focussing on three case-studies: the war in Ukraine, LGBTQI+ rights, and the 2024 European elections.

PROMPT uses the power of LLMs combined with dynamic network analysis for the in-depth recognition of formal patterns across media outlets, social networks and Wikipedia, in 8 languages, based on the rhetorical similarities observed qualitatively. With leading AI-scientists, academics and journalists, we’re developing a refined language model, a narrative monitoring dashboard and MOOCs/trainings for civil society activities.