The World as Viewed by 136 Countries: An Open, Curated, Annotated, and Multilingual Dataset of 118 Million State Media Articles

Key points

• State media outlets often project their governments' narratives to target both domestic and foreign audiences.1
• Because this content is plentiful and free to access, it appears widely in the web corpora that large language models (LLMs) are trained on and that retrieval systems draw from.1
• Using WorldStateMedia to audit these corpora, we find that, in English web crawls, the share of state media pages from editorially controlled (vs. independent) outlets rose from 39% in 2013 to 93% in 2025.1
• To characterize the differing narratives of news outlets, we compare Russian and Ukrainian state media in English and in their native languages, revealing asymmetric topical focus and opposing framings of contested issues such as Crimea.1

Sources (1)

• [1]The World as Viewed by 136 Countries: An Open, Curated, Annotated, and Multilingual Dataset of 118 Million State Media Articles
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 18, 03:49 AM

  In this paper, we introduce WorldStateMedia, an open dataset of 118 million state media articles from 136 countries in 114 languages.

  State media outlets often project their governments' narratives to target both domestic and foreign audiences.

Extractive summary: sentences quoted from the sources.