AI chatbot responses polluted by pro-Russian disinformation

Article excerpt

Researchers have discovered that popular AI chatbots are absorbing and amplifying pro-Russian disinformation, poisoning their responses to user queries. As millions turn to conversational agents for news and information, experts warn these systems lack robust defenses against coordinated propaganda campaigns. The finding reveals a fundamental vulnerability: chatbots train on internet data already contaminated by false narratives, then regurgitate that taint to millions of users seeking reliable answers. The scale of the problem remains unclear, but the pattern mirrors how disinformation spreads through social media, except now it arrives wrapped in the false authority of an AI system.