Study finds risky retail URLs in AI shopping answers
A Podqi study reviewed 6,795 shopping answers from Gemini, ChatGPT and Perplexity and found risky URLs in 4%, 8% and 15% of responses, respectively. Podqi defined risky sites as fake retailer impersonators, re-registered or inactive domains, and knockoff sellers; ChatGPT warned users about 24% of the risky sites it surfaced. Google spokesperson Jessica Levis Boord said Gemini performed best among the tested products, while OpenAI said it continues improving safeguards for potentially deceptive shopping results.
The cross-product comparison shows that AI-assisted shopping can direct users to fraudulent or otherwise unreliable websites, making source verification important before users submit personal or payment information.
AI chatbots are recommending fake retail websites, study finds
A study testing Gemini, Perplexity, and ChatGPT found that in as many as 15 percent of conversations, chatbots displayed ‘risky’ URLs.
A study testing Gemini, Perplexity, and ChatGPT found that in as many as 15 percent of conversations, chatbots displayed ‘risky’ URLs.
by Mia Sato
Nov 12, 2026, 5:00 PM UTC
