Abstract
Renaud Camus's 2011 book, Le Grand Remplacement, sparked the spread of the "Great Replacement" conspiracy theory in Western alt-right movements, particularly in the US and Europe. This theory has fueled violence, undermined democracy, and deepened political polarization, making it a crucial subject for study. Disinformation on social media promoting this theory offers valuable data for analysis, but the vast amount of multilingual content across platforms presents challenges. To address these, we introduce a hybrid AI research design that combines machine learning with Political Science methods to track the theory's evolution in English, French, Spanish, Portuguese, and Italian media. Our study first defines the semantic field of the "Great Replacement" across these languages, then applies traditional political discourse analysis to build an annotated corpus, identifying polarity in relevant sentences. We use AI to replicate this analysis with transformer-based algorithms and evaluate their performance against the annotated corpus.