Other

AI under the sight of the Russian Federation. Russian Pravda Network and global misinformation

Russian propaganda attacks artificial intelligence-a whole network called "Pravda" fruit millions of fakes, polluting information space, warns by international journalist Olga Tokaryuk. She explains how it works and calls Ukrainians for information hygiene. Did you know that Russia deliberately trains artificial intelligence models, such as Chat Gpt, on its propaganda? I guessed, but I recently learned the details.

It turns out that since 2022, Russia has developed a global network of sites in different languages, the so-called Pravda Network network, most of which have the word "Pravda" in their Internet address. These sites are more than 150 (40 of them are targeting Ukraine), they are administered from the occupied Crimea.

All of them do not produce any original content, but contain reposts in different languages ​​from Russian or pro-Russian telegrams, Russian state "media" and other Kremlin-related washing-in particular local, in other countries. The main topics of the news are Ukraine, the decline of the West, NATO and other typical propaganda narratives of Russia. Moreover, these sites are updated at impressive speed. Experts estimated that in 2024 only the Pravda network websites spread 3. 6 million articles.

At the same time, the audience of these sites is negligible, so the question arises - for whom does the Russian network produce this unrealistic amount of content? Researchers in several organizations that study Russian information operations - Newsguard, Viginum, DFRLAB and American Sunlight Project - have come to the conclusion that the purpose of the Pravda network is to train websters, large language models and generative artificial intelligence.

In simple words, all this huge number of Russian propaganda in different languages ​​is created and extended to make popular AI systems, such as Chat Gpt, Claude, Perplexity, Copilot, learned from this (DE) information and published it in their results. The News Guard study, on the example of the 10 largest companies that were "fed" with 15 Russian propaganda narratives, showed that in a third of cases, chatbots confirmed this content and referred to the Pravda network as a reliable source.