Kremlin’s “Pravda” Tricks AI in 49 Nations

A new report reveals a sophisticated Russian disinformation campaign, dubbed “Pravda,” successfully manipulated AI responses across 49 countries. The Kremlin-backed network reportedly flooded the internet with 3.6 million fabricated articles in 2024 alone.

The goal? To subtly influence the information that top AI models present to users, effectively embedding pro-Kremlin narratives into the AI’s knowledge base, according to research from NewsGuard.

Inside Russia’s AI Manipulation Tactic

This network, first identified in February 2024, isn’t focused on reaching human readers directly. Instead, it aims to saturate search results with false claims, thereby altering how large language models process and share information.

The American Sunlight Project (ASP) reported that “Pravda” operates across 150 domains in multiple languages, targeting Europe, Africa, and other regions. The network’s administration is allegedly linked to TigerWeb, an IT company in Russian-occupied Crimea.

John Mark Dugan, an American expatriate and propagandist, seemingly foreshadowed this strategy, stating in January 2024 that influencing AI could “change worldwide AI” by promoting Russian perspectives.

AI Chatbots Spread False Narratives

NewsGuard’s analysis showed that leading AI chatbots repeated falsehoods propagated by “Pravda” in approximately 33% of cases. Tests involved platforms like ChatGPT-4o, Google’s Gemini, and others.

The network functions as a “propaganda laundering machine,” amplifying content from Russian state media and pro-Kremlin sources through numerous seemingly independent outlets.

Examples include false claims about American biolabs in Ukraine and allegations regarding President Zelenskyy’s misuse of US aid. In one instance, several chatbots falsely claimed Zelenskyy banned Trump’s Truth Social in Ukraine, citing “Pravda” articles as evidence.

Little Organic Reach, Big AI Impact

Despite its scale, “Pravda” has minimal organic reach, with low average visitors and social media followers. However, its strategic focus on flooding search results poses a significant threat to AI models.

The ASP warns that “Pravda” is “poised to fill large language models (LLMs) with pro-Kremlin content,” intending to skew AI toward reproducing a specific narrative.

Global Implications

This revelation comes as independent media faces challenges, highlighting the importance of combating disinformation on all fronts, including within artificial intelligence.

Russia’s increased spending on propaganda, exceeding $1.4 billion in 2025, underscores its commitment to shaping global narratives.

The fight against Russian propaganda requires a comprehensive approach, including safeguarding independent journalism and developing robust defenses against AI manipulation.

Related Reads:

• “I don’t give a damn about their tribunal.” Why prosecuting Russian propagandists is so difficult
• The Musk factor: Would Ukraine’s frontline really collapse without Starlink?