Multilingual dataset helps detect russian propaganda in search results
SWARM: A Multilingual Human-Annotated Dataset for Russian Propaganda Detection in Search Engine Results
Computation and Language
Summary
Russian propaganda appears in many languages and on various websites, not just on known propaganda sources. The authors created SWARM, a dataset of search engine results in nine languages, each marked by experts for Russian propaganda content. They tested different ways to detect such propaganda and found that analyzing the content itself works better than just blocking certain sources. However, success depends on the model used, with some mistakes caused by confusing the topic of a page with support for propaganda.
What this means in practice
- •For search engine developers: Improve search results by detecting Russian propaganda across multiple languages using content-level analysis informed by the SWARM dataset.
- •For online content moderators: Use the dataset and evaluation methods to better identify and manage Russian propaganda appearing on mainstream and diverse web domains.
Authors
Manuel Tonneau, Abhinav Dubey, Farhan Shaikh, Ilaria Vitulano, Martha Stolze, Hale Dedeoglu, Clara Riechert, Ella Kuka, Maryna Sydorova, Mykola Makhortykh, Elizaveta Kuznetsova
Abstract
Russian state propaganda spreads across many languages and online spaces. Yet, most computational work examines only one such space, usually social media, in one or two languages, and analyses sources rather than content. We introduce SWARM (Search-Web documents Annotated for Russian propaganda, Multilingual), a dataset of 2,183 search engine results across nine languages and diverse web domains (e.g., news, blogs, government sites), each annotated by trained coders for whether it supports a recurring Russian propaganda narrative. We benchmark a source-based blocklist, supervised classifiers, and zero-shot LLMs against these labels. The blocklist misses most propaganda-supporting documents, because such content is not confined to flagged "propaganda" outlets but also appears on mainstream ones. Content-level analysis helps, though how much depends on the model: the strongest LLM reaches a positive-class F1 of 0.73, whereas the supervised classifiers reach only about 0.5, with the smaller LLMs over-predicting support, mistaking topical relevance for endorsement. Detecting search-borne propaganda thus requires per-language, content-level evaluation, which we hope SWARM and our evaluation code enable.