
AI chatbots are increasingly serving as primary information sources, yet they remain highly susceptible to sophisticated foreign disinformation campaigns. NewsGuard analyst Isis Blachez reveals that Russian-aligned networks, such as the "Pravda" site collective and the "Matryoshka" impersonation operation, are actively polluting the internet with millions of AI-generated articles to manipulate chatbot outputs. While Claude has historically been the most reliable model, recent audits show it is increasingly citing these fraudulent sources as credible, even when repeating debunked claims about the Ukraine war. These propagandists exploit the way AI models weigh information, effectively turning legitimate reporting on disinformation into a tool for lending credibility to falsehoods. This shift highlights a critical vulnerability in generative AI, where the sheer volume of synthetic content threatens to distort the truth and erode public trust in automated information systems.
Sign in to continue reading, translating and more.
Continue