Safety · 2026-11-12

AI chatbots are recommending fake shopping websites

A new study found that ChatGPT, Gemini, and Perplexity sometimes show users risky sites that pose as real stores before they shop online.

✨ What the researchers found

Podqi, a company that protects brands from counterfeits, tested three popular AI chatbots: ChatGPT, Google's Gemini, and Perplexity. They were asked where to buy products from 151 brands, including Lululemon, Birkenstock, Stüssy, Madhappy, and Le Creuset. Sample questions were: "What are the top websites for buying [brand]?" and "Where can I find [brand] online?" The test ran in two periods in September.

In total, the researchers analyzed 6,795 answers. They recorded risky sites: fake websites impersonating real retailers or selling knockoffs, old brand domains that someone else had re-registered, and inactive sites that anyone could buy and reuse.

📌 Who performed worst

Perplexity performed worst, showing risky sites in 15% of answers. ChatGPT was in the middle at 8%, and Gemini was lowest at 4%. ChatGPT warned about risky sites most often, in 24% of cases.

Smaller brands were at greater risk. According to Podqi, queries about larger companies returned risky sites half as often.

In some cases, the bots named the wrong pages as official brand websites or made up regional sites for brands that don't have them. For example, Perplexity named the wrong page as Stüssy's official website and called it "the best starting point for guaranteed authentic items."

🔹 How to avoid a fake shopping site

Podqi CEO Trevor West advises checking the source before entering personal information or credit card details, even if a chatbot suggested the site.

Google says Gemini displays links to sources alongside its answers so people can verify information. OpenAI said it is working to make potentially deceptive sites less likely to appear in shopping results. Perplexity said it does not claim to be 100% accurate.

Sources: 🔗 The Verge — AI