New study reveals AI chatbots may direct consumers to fraudulent retail sites

A recent investigation has uncovered a significant risk where consumers utilizing AI chatbots for product research or purchasing guidance could be steered toward deceptive or unsafe online vendors. The study, conducted by the IP protection firm Podqi and shared with The Verge, tested major language models including ChatGPT, Gemini, and Perplexity. The researchers posed shopping-related queries to the AI systems, focusing on 151 different brands. The questions included inquiries such as where to find specific brands online, what the top purchasing websites were for those brands, and whether the companies offered international shipping options. These tests were carried out across two separate testing periods during September.