AI for Political Analysis: Disinformation and the “Great Replacement” Theory

• In Event: Emerging Scholars Lightning Round: Research Design

Abstract

The release of Renaud Camus's "Le Grand Remplacement" (The Great Replacement) in 2011 in France is recognized as the initial catalyst for the proliferation of this conspiracy theory within Western alt-right movements, notably in the US and Europe. This theory has been identified as a significant motivating factor behind instances of extreme violence, posing a threat to democracy and exacerbating political polarization. The proven disruptive potential of this conspiracy theory underscores its importance as a phenomenon that deserves thorough examination.

The proliferation of disinformation in social media supporting that idea of national populations being replaced by immigrants provides an useful data source to enable evidence-based analysis of this phenomenon. However, the amount of information, the co-existance of different social media platforms, and the multilingual environment of European politics introduces severe limitations to perform systematic and generalizable inference. Our research aims to tackle those limitations, introducing a hybrid Artificial Intelligence research design, where computational techniques are added to the toolbox of Political Science to enable political discourse across multiple languages. Recognizing AI's dual role in both spreading and potentially preventing disinformation, we utilize state-of-the-art machine learning algorithms and advanced natural language processing techniques to train a machine learning model able to trace the theory's evolution in English, French, Spanish, Portuguese, and Italian media discourses.

With this purpose, we start defining the semantic field addressing the 'Great Replacement Theory' across the four languages. With this purpose, we explore linguistic and contextual variances in multilingual expressions using computational linguistics techniques to derive terms consistent with the political science literature on this topic. Second, we apply conventional political discourse analysis techniques to produce an annotated corpus where we identify the polarity of relevant sentences regarding those topics more closely connected to the 'Great Replacement' theory. Third, we perform the same polarity analysis, this time using several transformer-based Artificial Intelligence algorithms, using the semantic set as an ontology consistently with computer science standars. Finally, using the annotated corpus as a gold standard, we evaluate the performance of the tested AI approaches, gaining insights about their potentials and limitations.


Our research sheds light on the mechanics of narrative spread and evolution in the case of the "Great Replacement" theory, contributing to tackle the impact of disinformation in democratic societies. In addition to that, we contribute with a research design that can be applied to other research topics, particularly in the field of political communication, discourse analysis or disinformation studies. In order to ensure that our approach is reusable, we make our code and data available through the Harvard Dataverse repository, so researchers can independently evaluate our analysis, and use the code for their own research in a way that ensures a strong commitment to ethical research practices, particularly concerning data access.

Authors

• Erik Marino, University of Évora

• Suso B Baleato, Universidade de Santiago de Compostela

• Ana Sofia Ribeiro, CIDEHUS - University of Évora