AI Found Me A Fake Store

The real news

The story, straight

A shop's web address is its online front door. A chatbot naming that door doesn't prove the brand is behind it.

Podqi, which sells brand protection, tested ChatGPT, Gemini and Perplexity using its own shopping questions. As reported by The Verge, Podqi found 718 risky websites among the addresses their answers named.

These included apparent brand impersonators, suspected counterfeit sellers and inactive sites. Inactive doesn't mean fraudulent: it means nobody currently operates the site.

The uncomfortable bit: the answers sometimes recommended risky addresses as places to shop. The Verge reports that Perplexity called the wrong page Stüssy's official website and described it as a starting point for authentic items.

My AI stylist found streetwear. Unfortunately, the store was wearing a disguise.

Chatbots build answers from learned text and search results. An old address can survive in that information after a shop closes or ownership changes.

A fluent recommendation isn't an ownership check.

This measured answers, not shoppers losing money. And Podqi sells a service addressing the problem it studied, so treat its findings as a vendor's test, not an independent audit.

The useful move: reach regional stores through the brand's main website. Check the address before entering personal or payment information.

A chatbot can help find a hoodie; don't let confidence become its security certificate.

Source: The Verge; Podqi's AI search brand-risk report.

Repost before your shopping assistant recommends a fitting room made of fog.

Thanks for supporting Lil'meme. Your support is appreciated.