Large-scale detection of multilingual coordinated activity on Telegram
摘要
By committing to user privacy, resisting collaboration with law enforcement, and maintaining minimal moderation, Telegram has become a haven for extreme opinions and a fertile ground for information operations (IOs). In this study, we analyzed over four million political messages in English and Russian shared on Telegram around the 2024 U.S. Presidential Election, a period of intensified political activity worldwide, to uncover orchestrated IOs. Our multilingual investigations, based on coordinated inauthentic behavior (CIB) detection techniques and Telegram-specific analytics, uncovered five IOs tied to diverse political interests across the globe, including a Russia-backed influence campaign and a pro-Palestine fully automated amplification network. These findings underscore the need for stronger policies to detect and moderate coordinated inauthentic behavior on Telegram and better safeguard its organic discourse.