Researchers claim that at a time when the US is reportedly pausing its cyber operations against Moscow, a vast Russian misinformation network is influencing Western AI chatbots to produce pro-Kremlin propaganda.

It is claimed that the Pravda network, a well-funded Moscow-based organization that disseminates pro-Russian narratives around the world, is skewing chatbot results by overloading large language models (LLM) with anti-Kremlin propaganda. The misinformation watchdog NewsGuard conducted a survey of ten top AI chatbots and discovered that they promoted a pro-Moscow agenda by repeating lies from the Pravda network over 33% of the time.

According to the findings, the threat extends beyond generative AI models that detect misinformation on the internet and includes the intentional targeting of chatbots to reach a larger audience in a manipulation technique known as LLM grooming. In a paper, NewsGuard researchers McKenzie Sadeghi and Isis Blachez stated that massive amounts of Russian propaganda -- 3,600,000 articles in 2024 -- are now incorporated into the outputs of Western AI systems, infecting their responses with false claims and propaganda.

The nonprofit American Sunlight Project issued a warning in different research about the Pravda network's expanding reach, sometimes referred to as "Portal Kombat," and the possibility that its pro-Russian material was overflowing with huge language models' training data. Experts cautioned that without regulation in the US, this misinformation would spread more widely.

President Donald Trump has been pressing for talks to end the three-year war in Ukraine, and the reported move comes days after the US leader chastised his Ukrainian counterpart, Volodymyr Zelensky, in a heated White House meeting. The Pentagon declined to comment on the reports.