AI Chatbots Surface Fake and Imposter Retail Websites in Up to 15 Percent of Shopping Queries, Study Shows

Major Chatbots Direct Shoppers to Fake and Risky Sites

As consumers increasingly rely on artificial intelligence tools to discover products and compare prices, popular AI chatbots are directing users toward imposter storefronts and suspicious web domains.

A study conducted by intellectual property protection firm Podqi found that leading AI engines—including OpenAI’s ChatGPT, Google’s Gemini, and Perplexity—frequently surface "risky" URLs when prompted with e-commerce questions. Depending on the chatbot, risky links appeared in up to 15 percent of responses.

Podqi defines risky sites as counterfeit domains impersonating legitimate retailers, websites selling knockoffs, expired brand domains reregistered by third parties, or inactive web pages susceptible to domain takeover.

Perplexity Registers Highest Risk Rate Among Tested Chatbots

The testing evaluated a total of 6,795 generated answers across two testing periods in September. Podqi prompted the models using a selection of 151 small and large companies, including fashion labels Lululemon, Birkenstock, Stüssy, and Madhappy, alongside cookware manufacturer Le Creuset.

The queries tested common shopping interactions, such as asking where to buy specific brands online, whether brands ship internationally, or what official regional websites exist.

Perplexity performed the worst among the three platforms, linking to or mentioning risky sites in 15 percent of its responses. In one notable instance, Perplexity surfaced an incorrect link as Stüssy’s official website, identifying it as "the best starting point for guaranteed authentic items."

ChatGPT generated risky links in 8 percent of its answers. However, OpenAI's tool performed best at warning users about suspicious links, providing explicit cautions in 24 percent of the instances where a risky domain was displayed.

Google’s Gemini achieved the best overall performance, citing or displaying risky sites in 4 percent of its answers.

Smaller Brands and Regional Queries Face Higher Exposure

The study highlighted specific vulnerabilities based on brand size and query structure. AI chatbots surfaced deceptive sites about half as often when answering questions about major corporate brands compared to smaller, niche businesses.

Queries asking for regional brand websites proved particularly prone to errors. Chatbots regularly displayed fake regional storefronts, including generating fraudulent regional URLs for brands that do not operate separate regional websites.

"Part of this piece, too, is making sure that the consumers know, even though you are being fed a result in a particular chatbot, it is always best to verify the source of that before entering personal information or credit card information," said Podqi CEO and co-founder Trevor West.

Tech Companies Respond to Safeguard Concerns

In response to the findings, representatives from the respective AI companies outlined their efforts to address source reliability and domain verification.

Google spokesperson Jessica Levis Boord stated: "We take information quality seriously, and are happy to see that Gemini performed best among the products tested. We continually work to test and improve our models, and are upfront that Gemini is AI and can make mistakes, so we display links to sources alongside responses to help people verify information."

OpenAI spokesperson Taya Christianson reiterated the company's commitment to consumer safety: "Our goal is for ChatGPT to connect people with reliable places to shop. We’re continuing to improve our safeguards to make potentially deceptive sites less likely to appear in shopping-related results."

Perplexity spokesperson Jesse Dwyer defended the platform's core focus on information retrieval: "Perplexity is consistently the most accurate AI on credible research benchmarks, and has been for four years. We never claim to be 100% accurate, but we do claim to be the only company relentlessly focused on it."