Russia has quietly hijacked Wikipedia, injecting pro‑Kremlin edits that reshape public understanding of the Ukraine war. Analysts trace a coordinated effort that began after the 2022 invasion, using crowd‑sourced pages to push sanitized narratives. This manipulation extends to popular AI chatbots, which pull from the same skewed articles.

The Pravda network, a web of 193 sites identified by France’s VIGINUM watchdog, republishes content from Russian state media and Telegram channels without original research. In December 2023, a Wikipedia entry on the Red Sea was edited to cite pravda‑fr.com, a move flagged as foreign digital interference. Such inserts demonstrate a systematic laundering of Kremlin tropes into neutral reference material.

The Atlantic Council researchers mapped how these rewritten pages feed large‑language models. By supplying LLMs with biased Wikipedia snapshots, the Kremlin embeds pro‑Russian framing into chatbots that users worldwide trust for quick answers. The study cites elections in Romania and Moldova as test cases where the network amplified Kremlin narratives, showing the reach of this digital disinformation.

This coordinated poisoning of knowledge platforms undermines factual discourse and skews AI outputs. The scale of the operation—spanning 80 countries—shows that state actors can weaponize encyclopedic content as cheaply as they can launch cyberattacks. Detecting and countering these edits is essential if digital literacy and trustworthy AI are to survive.