An Avalanche of Images on Telegram Preceded Russia's Full-Scale Invasion of Ukraine

TL;DR

The paper addresses how visually-driven propaganda influences geopolitical instability by analyzing Politically Salient Image Patterns (PSIPs) on Telegram surrounding Russia's 2022 invasion. It combines unsupervised image clustering (MobileNet features, FAISS, spectral clustering) with expert coding to identify PSIPs and manipulated imagery from 989 milbloggers, revealing large pre-invasion spikes in posts and images ($8,925\%$, $5,352\%$; $p<0.001$) and a rise in manipulated content. It finds PSIPs operate through ingroup solidarity, outgroup vulnerability, and epistemic insecurity, and documents numerous manipulated exemplars across multiple languages. The work demonstrates that AI-assisted, large-scale analysis of visual media can illuminate the dynamics of political instability and informs calls for real-time multi-modal monitoring to support peace and conflict research.

Abstract

Governments use propaganda, including through visual content -- or Politically Salient Image Patterns (PSIP) -- on social media, to influence and manipulate public opinion. In the present work, we collected Telegram post-history of from 989 Russian milbloggers to better understand the social and political narratives that circulated online in the months surrounding Russia's 2022 full-scale invasion of Ukraine. Overall, we found an 8,925% increase (p<0.001) in the number of posts and a 5,352% increase (p<0.001) in the number of images posted by these accounts in the two weeks prior to the invasion. We also observed a similar increase in the number and intensity of politically salient manipulated images that circulated on Telegram. Although this paper does not evaluate malice or coordination in these activities, we do conclude with a call for further research into the role that manipulated visual media has in the lead-up to instability events and armed conflict.

Figures (7)

• Figure 1: Raw number of collected Telegram posts and their associated images from 2015 to 2023.
• Figure 2: Month-to-month difference in total number of posts and posts containing images. A spike in activity corresponds with the beginning of Russia's full-scale invasion in the Ukraine.
• Figure 3: Politically salient image exemplars at the time of the Russian full-scale invasion of Ukraine.
• Figure 4: Politically salient images at the time of Russia's full-scale invasion of Ukraine that are fake or misleadingly out of context.
• Figure 5: (A) Pro-Russian meme mocking anti-Putin journalist Arkady Babchenko. The text on the shirt was inserted into the photo. Babchenko's head appears to be inserted onto someone else's body. (B) JPEG blocking artifacts from alterations of the image; (C) Median Filtering results reveal alterations of the image. Media forensics tools can detect the tampering lin2009fast, but cannot understand its political relevance.