AI Chatbots Recommend Fraudulent Retail Sites to Shoppers in Podqi Study

Leading AI platforms—including Perplexity, ChatGPT, and Google Gemini—routinely surface counterfeit storefronts and re-registered defunct domains when queried for retail purchasing links.

* Search algorithms remain vulnerable to domain manipulation. A study of 6,795 chatbot responses across 151 brands by Podqi found that models frequently ingest and prioritize malicious websites optimized specifically to deceive web crawlers. * Warning systems largely fail to trigger. When chatbots surfaced risky retail links, 57 percent of the responses actively endorsed the fraudulent domains for purchases, while only 9 percent provided caution flags to the user. * Boutique and mid-tier labels take the brunt. Smaller commercial brands faced approximately double the rate of exposure to impersonator sites compared to major global players like Lululemon or Birkenstock.

Until frontier lab developers integrate authenticated merchant registries into real-time search pipelines, direct URL verification remains a non-negotiable step before checkout.