A Russian disinformation network is using AI chatbots to disseminate propaganda and fake news online through AI answers, a recent NewsGuard report has found. The study reveals that pro-Kremlin propaganda is distorting artificial intelligence’s work, impacting millions of people worldwide.
According to the research, the primary driver of this campaign is the Moscow-based disinformation network Pravda. By overloading search engines and web crawlers with a lot of disinformation, it aims to affect the answers of AI chatbots rather than directly persuading humans.
A large-scale campaign is being pursued by a Moscow-based disinformation network called “Pravda” (which has a dedicated website and a local language for nearly every nation in the EU). It purposefully infiltrates the data that artificial intelligence chatbots retrieve and publishes propaganda and false claims to influence the responses of AI models on news topics, NewsGuard has confirmed.
Russian Pravda network floods AI chatbots with millions of articles – NewsGuard
According to the study, more than 3.6 million articles published by the Russian network Pravda in 2023 were used by well-known AI chatbots from Microsoft, Google, OpenAI, xAI, Meta, Anthropic, Mistral, Perplexity, and You.com. The network has disseminated at least 207 verified fakes, most of which have some connection to Ukraine.
Unlike disinformation campaigns on well-known propaganda media that aim to reach human audiences, the Pravda network appears to design itself to influence AI chatbots. The network makes sure that AI algorithms trained on publicly accessible data take in and replicate Kremlin-backed misinformation by overloading search engines with fake stories and allegations.
Instead of producing original content, the Pravda network republishes news from Russian state propaganda media, pro-Kremlin influencers, and government organizations to compile and spread Kremlin narratives. This enables AI models, trained on vast online datasets, to incorporate these stories and mistakenly portray them as fact.
33% answers from AI chatbots reproduced false reports from Russia’s Pravda network
The network has distorted how big language models interpret and display news and information by overloading web crawlers and search results with pro-Kremlin disinformation.
When American pro-Russian fugitive propagandist John Mark Dougan spoke to a group of Russian officials in Moscow last January, he hinted at the infection of Western chatbots by saying, “By pushing these Russian narratives from the Russian perspective, we can actually change worldwide AI.”
According to a NewsGuard assessment, the top AI chatbots 33 percent of the time reproduced false reports that were laundered by the Pravda network. The pro-Kremlin campaign changes training datasets that use large language models (LLMs) by controlling search queries. This changes how AI summarizes news and responds to user questions.
NewsGuard’s findings confirm a February 2025 report by the American Sunlight Project, which warned that Russia’s Pravda network was likely designed to manipulate AI models rather than to generate human traffic. As we have found in our previous research, these websites have low-volume traffic.
“The long-term political, social, and technological risks linked to potential LLM grooming within this network are high. The larger a set of pro-Russia narratives is, the more likely it is to be integrated into an LLM,” the ASP concluded.
Russia launched Pravda disinformation websites in dozens of languages
Following Russia’s all-out war against Ukraine on February 24, 2022, the Pravda network launched in April 2022. It was first identified in February 2024 by Viginum, a French government watchdog that monitors foreign meddling. Since then, the Russian disinformation network has grown, targeting now at least 49 countries in dozens of languages across 150 websites.
The network of websites was named “Portal Kombat” by the Viginum Agency report, and one of its main goals was said to be the spreading of pro-Russian disinformation and propaganda in different EU languages—in particular, French, German, Spanish, and English—through a web of copycat websites named Pravda* that were created in 2023 and became active in the second half of the year.
From EDMO investigations it came out that in the span of one week (20-26 March 2024), the Pravda network was activated in 19 EU countries: Greece, Italy, the Netherlands, Denmark, Croatia, the Czech Republic, Slovakia, Bulgaria, Romania, Finland, Sweden, Portugal, Estonia, Latvia, Lithuania, Hungary, Cyprus, Ireland, and Slovenia. (Austria, Belgium, Malta, and Luxembourg are not covered directly, but language-wise they are covered by other outlets.).
During the same period (20-26 March 2024), Pravda websites were also activated in non-EU member states in Europe—Norway, Moldova, Bosnia, Albania, and North Macedonia—and even outside Europe, in the République centrafricaine (CAR), Niger, Taiwan, and Japan.
List of Pravda websites
- Pravda-de (German) – 22 February 2023
- Pravda-fr (French) – 24 June 2023
- Pravda-pl (Polish) – 24 June 2023
- Pravda-es (Spanish) – 24 June 2023
- Pravda-en (English) – 24 June 2023
- Pravda-it (Italian) – 20 March 2024
- Pravda-gr (Greek) – 20 March 2024
- Pravda-dk (Danish) – 20 March 2024
- Pravda-nl (Dutch) – 20 March 2024
- Pravda-hr (Croatian) – 20 March 2024
- Pravda-cz (Czech) – 20 March 2024
- Pravda-sk (Slovak) – 20 March 2024
- Pravda-bg (Bulgarian) – 20 March 2024
- Pravda-ro (Romanian) – 20 March 2024
- Pravda-fi (Finnish) – 20 March 2024
- Pravda-se (Swedish) – 20 March 2024
- Pravda-pt (Portuguese) – 20 March 2024
- Pravda-ee (Estonian) – 20 March 2024
- Pravda-lv (Latvian) – 20 March 2024
- Pravda-lt (Lithuanian) – 20 March 2024
- Pravda-hu (Hungarian) – 20 March 2024
- Pravda-ie (Irish) – 26 March 2024
- Pravda-si (Sloven) – 26 March 2024
- Pravda-cy (Cyprus) – 26 March 2024
- Pravda-no (Norway) – 20 March 2024
- Pravda-md (Moldavia) – 20 March 2024
- Pravda-ba (Bosnia) – 26 March 2024
- Pravda-al (Albania) – 26 March 2024
- Pravda-mk (North Macedonia) – 26 March 2024
- Pravda-ne (Niger) – 26 March 2024
- Pravda-cf (République centrafricaine)
- Pravda-jp (Japanese) – 26 March 2024
- Pravda-tw (Taiwanese) – 26 March 2024
This attribute indicates that the Pravda network of websites uses a high level of automation and little to no human moderation. This makes it highly effective in publishing a large volume of content and very cost-effective, even if the quality of translation is poor in some languages, EDMO wrote.
Of the 150 websites in the Pravda network, nearly 40 are Russian-language sites publishing under domain names targeting specific cities and regions of Ukraine, including News-Kiev.ru, Kherson-News.ru, and Donetsk-News.ru. We have identified them in 2024 in our analysis Russian network of regional propaganda outlets in Ukraine disclosed.
According to Viginum, the Pravda network is run by TigerWeb, an IT firm based in Russian-occupied Crimea. TigerWeb is owned by Yevgeny Shevchenko, a Crimean-born programmer who formerly worked for Krymtechnologii, a company that built sites for the Russian-backed Crimean government.
In total, 56 out of 450 chatbot-generated responses included links to stories spreading false news published by the Pravda network. The chatbots used 92 different articles from the network that spread falsehoods. Two of them used as many as 27 Pravda articles from domains in the network, such as Denmark.news-pravda.com, Trump.news-pravda.com, and NATO.news-pravda.com.
Russian Pravda websites target AI tools and search engines
Despite its scale and size, the Russian disinformation network receives little to no organic reach. According to SimilarWeb data, Pravda-en.com, an English-language site within the network, has an average of less than one thousand monthly unique visitors, and it’s an invisible website compared to the 74 million estimated monthly visitors to Russian state-run RT.com.
But this small coverage masks the network’s wider influence. These disinformation platforms seem to focus on saturating search results and AI web crawlers with automated content at scale, rather than creating a human audience across social media as RT and Sputnik typically do.
In addition, a January 2025 report from Google said it observed that foreign actors, including Russians, are increasingly using AI and search engine optimization to make their disinformation and propaganda more visible in search results.
One Russian government-backed group used Gemini to request help with many tasks, including rewriting publicly available malware into another language, adding encryption functionality to code, and explaining how a specific block of publicly available malicious code functions, Google reported.
Will AI companies identify and combat disinformation networks?
Experts caution that AI-generated responses will continue to be used as instruments of foreign propaganda unless AI companies take proactive measures to identify and combat the influence of disinformation networks. Improved filtering systems, more openness in AI training data, and collaborations with impartial fact-checkers to validate AI-generated content are some possible remedies.
The Pravda network’s approach may have long-term effects on global information security, impacting political processes, public opinion, and international relations as AI systems are increasingly utilized to influence public opinion.
Policymakers, AI developers, and the general public face a significant dilemma as generative AI’s vulnerability to misinformation grows in importance as a source of information. If we don’t take action, AI might inadvertently propagate state-sponsored propaganda and alter global discourse to align with authoritarian objectives.
