Fraudsters Use Fake Websites to Exploit ChatGPT Recommendations
Scammers are creating cloned retail websites that appear in ChatGPT search results, tricking buyers into handing over money and banking details for products that never arrive. Ask Silver, a scam-checking service, has identified fraudulent sites impersonating Russell & Bromley and Dunelm in the AI tool's responses.
The scheme works because generative AI and LLM systems learn from web content, including pages created by fraudsters. When malicious content enters the training data-a process called "poisoning"-the AI can unknowingly recommend fake stores to users.
How the Scam Operates
A user asks ChatGPT for shopping recommendations. The tool returns product listings with prices and links that appear legitimate. The cloned websites look credible, often featuring steep discounts and URLs similar to official retailers.
Russell & Bromley became a target after the brand went into administration in January 2026 and was absorbed by Next. No official Russell & Bromley website exists anymore, but customers still search for it by name. Fake sites identified include therussellbromleyofficial, russellandbromleylondon, and russellbromleyonlineuk.
Once a buyer completes a purchase, the fraudster keeps the money and customer banking data vanishes.
Red Flags to Watch
- Domain names with extra words like "official" or "deals"
- Non-standard domain extensions (.co.uk and .com are standard for UK retailers)
- Payment by bank transfer only-legitimate retailers offer multiple options
- Discounts of 80% or more on items
What to Do
Go directly to retailer websites rather than clicking links from AI-generated results. Dunelm directs customers to www.dunelm.com or its official app. Russell & Bromley operates through the Next website.
If you've already provided financial details to a fraudulent site, report it to your bank and to Report Fraud immediately.
Louise Baxter, head of the scams team at National Trading Standards, said: "Consumers are increasingly turning to AI tools for advice and recommendations, but criminals are adapting just as quickly. The fact that scam websites can appear in AI-generated results is worrying, and is a stark reminder that fraudsters will exploit any new technology that helps them reach potential victims."
ChatGPT removed the identified fraudulent websites from its search index. Users can report policy-violating sites through the platform's reporting form.
Your membership also unlocks:
