Recruit and threaten: hate speech detection within the pro-Wagner digital ecosystem on Telegram

The PMC Wagner emerged as a central actor in the Russian power projection. This research consists of an evidence-based analysis to grasp the relationships within the pro-Wagner digital ecosystem. Through Socio-Semantic Network Analysis and hate-speech detection using a machine learning model, the study aims to map the PMC Wagner on Telegram and reconstruct its morphological characteristics. The study demonstrates that some intermediaries are essential in spreading propaganda due to a combination of their position and the violence of their discourse. Targeting the videos of the most relevant intermediaries is an efficient strategy to prevent other violent extremist actors.