Skip to main content
Promotional banner for the pentest readiness checklist
Test findingAI acted on its ownSeverity N: legal or test findingReported: press reporting only

Podqi study found ChatGPT, Gemini and Perplexity showed shoppers fake and risky retail websites

IP protection firm Podqi prompted ChatGPT, Gemini and Perplexity with shopping questions about 151 brands and analysed 6,795 answers. All three cited or showed risky sites, including fake sites impersonating retailers, reregistered old brand domains and dead sites. Perplexity showed risky sites in 15 percent of answers, ChatGPT in 8 percent and Gemini in 4 percent, and smaller brands were more affected. The study used crafted prompts and reported no actual consumer losses; Google and OpenAI said they keep improving, while Perplexity defended its accuracy.

What the AI did

When asked shopping questions about 151 brands, ChatGPT, Gemini and Perplexity linked to or mentioned risky websites in some of their answers, including fake sites posing as retailers, old brand web addresses now owned by someone else, and dead sites. In one conversation, Perplexity named the wrong page as Stüssy's official website.

First reported November 12, 2026 · Added to the register November 17, 2026 · 1 source

Model
ChatGPT, Gemini, Perplexity
Affected
Stüssy
When it happened
Not stated in the sources
First reported
November 12, 2026

What this means for you

Could this affect you?

Yes

If your brand sells online or your staff use AI chatbots to find where to buy. Shoppers using AI chatbots, and brands (especially smaller ones) whose customers may be steered to imposter or reregistered web addresses, were exposed, although the study used crafted questions and reported no actual consumer losses.

What to check

  • Monitor how AI chatbots cite your official and regional web addresses.
  • Recover or watch web addresses your brand has let lapse.
  • Verify a website is genuine before entering personal or payment details.
  • Brief procurement teams to check sites suggested by AI chatbots before buying.
Every fact and its source (8)
  1. Brands covered151
    “involving a selection of 151 brands”[1]
  2. Answers analysed6,795
    “Over the course of 6,795 answers”[1]
  3. Perplexity risky-site rate15 percent of answers
    “Perplexity performed the worst, showing risky sites in 15 percent of answers”[1]
  4. ChatGPT risky-site rate8 percent
    “ChatGPT fell somewhere in the middle at 8 percent”[1]
  5. Gemini risky-site rate4 percent of answers
    “linking to or mentioning a risky site in 4 percent of answers”[1]
  6. ChatGPT warning rate on risky sites24 percent
    “doing so 24 percent of the time”[1]
  7. OpenAI responseImproving safeguards against deceptive sites
    “We’re continuing to improve our safeguards to make potentially deceptive sites less likely to appear in shopping-related results”[1]
  8. TesterPodqi, IP protection firm
    “The study conducted by IP protection firm Podqi”[1]

Sources

  1. AI chatbots are recommending fake retail websites, study finds
    theverge.com · November 12, 2026

How this record is classified. Severity N: a legal action or test finding, not a harm. OECD level: hazard, an event that could plausibly have led to harm. Evidence: Reported, meaning press reporting only.

The facts in At a glance and in Every fact and its source are quoted from the sources listed. The summary, What the AI did and What this means for you are written from those sources and checked against them automatically before publication. Records reflect what has been disclosed, not everything that has happened. Incidents are found through news feeds and the GDELT Project.

Promotional banner graphic asking if you are ready for PCI DSS 4.0 with a call-to-action to get the guide