/ai-search554 words

AI-Cloned Fake Stores Flood Search Results Ahead of Black Friday

Norton expects a significant rise in AI-created fake storefronts before Black Friday on 27 November, with criminals 'poisoning' ChatGPT so it recommends scam sites.

Unreal deals: the scam websites targeting Black Friday bargain hunters
Unreal deals: the scam websites targeting Black Friday bargain huntersathree23 / pixabay
  • Norton expects suspect sites to rise significantly between Black Friday (27 November) and Christmas
  • National Trading States says criminals are 'poisoning' ChatGPT so it recommends scam websites
  • Clone sites imitating Russell & Bromley used URLs like therussellbromleyofficial, documented in June
  • Fake stores offer discounts up to 80% and push bank transfer or cryptocurrency payments
  • The Co-operative Bank launched a tool that checks suspicious texts, emails and websites for safety

Cybersecurity firms expect a significant rise in AI-generated fake storefronts between Black Friday on 27 November and Christmas, with clone sites surfacing both in regular search results and in answers produced by AI chatbots. National Trading Standards has stated that criminals are "poisoning" AI models such as ChatGPT so the chatbots recommend scam websites to users — a confirmed statement from the UK consumer protection body, not tool speculation.

The scale follows a yearly pattern. "The increase is clear every year at this time between Black Friday and Christmas," said Luis Corrons, a security researcher at antivirus company Norton. Norton says it expects the number of suspect sites to rise significantly as the gift-buying season accelerates.

How do fake storefronts reach shoppers?

The clones mimic legitimate brand sites down to the URL. In June, the Guardian documented websites imitating the shoe seller Russell & Bromley using addresses such as therussellbromleyofficial and russellbromleyonlineuk. The sites carry the brand name in the domain, list products in the correct colours, and attach steep discounts — 80% in some cases — to pull buyers in before payment.

The distribution channels mirror ordinary e-commerce marketing. "They reach [AI] agents the same way they reach people: through populating search results, Google ads and other paid ads, compromised sites and social posts pushing a hot sale," said Michael Vainshtein, co-founder of security company Guardio. In other words, the same manipulations that push a site up organic and paid results also feed the outputs of AI assistants that summarize those results.

Norton's global head of scam research, Leyla Bilge, flagged one demographic in particular. Younger shoppers are more likely to encounter scams through social media and to buy on impulse, she said. "If a hard-to-find item suddenly turns up cheap from a seller you've never heard of, treat that as a warning sign rather than a stroke of luck," Bilge advised.

What marks a cloned store?

Several signals recur across these AI-built sites, according to the warnings:

  • Web addresses that embed the brand name plus extra words such as "official", "online" or "deals" instead of a clean .co.uk or .com domain.
  • Discounts that reach 80% in some cases on in-demand items.
  • Payment pages that push bank transfer or cryptocurrency instead of standard card processing.
  • PayPal offered only through the "send to a friend" option, presented as a way to avoid fees.

The payment route is the sharpest tell. An immediate red flag, the guidance says, is a store that accepts only bank transfer or cryptocurrency — methods that offer little recourse once money moves.

What can shoppers do?

The Co-operative Bank has launched a tool that checks suspicious texts, emails and websites and reports whether they are safe. Shoppers who have already handed over financial details should contact their bank and file a report with Report Fraud. Sites that violate ChatGPT's policies can be reported through OpenAI's dedicated form.

For search professionals, the story carries a two-sided lesson: the same search and paid-listing surfaces that legitimate retailers depend on for seasonal traffic are the attack vector for clones, and AI assistants now inherit whatever those surfaces return.

Monitor Norton's threat data and National Trading Standards advisories through the January sales period, when the clone-site surge historically persists past Christmas.

via theguardian.com (Source)

More from Nathan Brooks

Nathan Brooks

Show full bio

Senior reporter covering companies, deals and strategy in Search rankings and algorithm updates, generative engine optimization at SERP Journal.

228 articles