AI Chatbots Direct Users to Fraudulent Online Stores
November 16, 2026 · 1 min read
Also published in Deutsch, Nederlands, Português (Brasil), Norsk, Polski
A recent study by Podqi reveals that artificial intelligence services including Chat GPT, Gemini, and Perplexity risk directing consumers toward fake online stores and websites. The investigation involved asking the three AI services questions about one hundred fifty-one brands. A total of six thousand seven hundred ninety-five responses were analyzed during two test periods in September. The results indicate that all services referred users to risky websites. Perplexity performed the worst, mentioning or linking to such sites in fifteen percent of responses.
The corresponding share for Chat GPT was eight percent, while Gemini recorded four percent. The study defines risky websites as including fake online stores posing as established brands, sites selling pirated copies, and old domains taken over by other actors. The issue was particularly prevalent regarding questions about brands' regional websites. Chat GPT was best at warning users about risks, but did so in only twenty-four percent of cases when a risky website appeared.