Mapping social and political spaces on digital platforms and AI systems

A large number of works take interest in the role of social media platforms – including their affordances, particular culture, and algorithmic mediation – on phenomena such as the segregation of political communities, polarization, and lack of diversity of news diets. Behind this new field of studies, an important methodological lever is the political characterization of users and content into Left-Right or Liberal-Conservative scales and communities. In this talk, we will review the work of the European Polarisation Observatory (EPO), an initiative led by Sciences Po and CNRS together with seven additional partners in Europe.

One of the main activities of the EPO is the production of populations of social media users in different countries, mapped onto several comparable ideology and issue dimensions: Left-Right, attitudes towards the EU, immigration, environmental protection, or trust in elites and institutions, to name a few. These populations, mapped along the dimensions that are relevant for different case studies in different national contexts, constitute the basis for different works carried out at EPO. We will illustrate some of these works with studies regarding the political dimensionality related to the spread of misinformation online, the cartography of political competition online, and the mapping of profiles learned by algorithms computing recommendation in social media.

Pedro Ramaciotti is the Chair of AI in Social Sciences and Humanities at the French National Research Center CNRS. He’s affiliated with the Complex Systems Institute of Paris and the médialab at Sciences Po. He’s also a Research Fellow at the Learning Planet Institute from University  Cergy Pointoise.