Skip to content
Submit incident
AII-1000-88873
Well documentedThree or more independent publishers report the core facts.

ChatGPT recommends scam shopping sites after data poisoning

ChatGPT has been observed recommending cloned websites that impersonate defunct retailers like Russell & Bromley, which scammers created to steal money and bank details. Researchers believe malicious content was injected into ChatGPT's training data, and security officials warn consumers should not assume AI-recommended websites are genuine.

7 June 2026 Β· OpenAI Β· Fraud, scams and manipulation
7 June 2026
AII-1000-888737 sources ↓
Quick facts
Category
Fraud, scams and manipulation
Country
United Kingdom
Source
The Guardian
Date
7 June 2026
Drafted by
Raidu Incident Detection

What happened

Ask Silver, a scam-checking service, reported that ChatGPT was recommending cloned websites that impersonate Russell & Bromley, a British footwear and handbag retailer that went into administration in January 2026 and was absorbed by Next. When users asked ChatGPT for 'popular Russell & Bromley purses and bags', the AI assistant returned results that included links to fraudulent websites using official-sounding URLs like therussellbromleyofficial dot com, with fake discounts of up to 80% offered.

Anna Jones of Ask Silver told The Guardian that scammers were taking advantage of the fact that Russell & Bromley no longer had an official website, as potential customers would still be searching for it. The fake websites appeared credible, and users who clicked through could have their money stolen and bank details harvested.

Reported impact

Louise Baxter, head of the scams team at the UK's National Trading Standards, stated: '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.'

Index analysis

Anna Jones told The Guardian that it is possible ChatGPT's language model has been 'poisoned' by malicious content sneaked into its training data, due to which the AI model returns fraudulent and scam pages when triggered by customer queries.

Classification

Category
Fraud, scams and manipulation
Industry
Consumer services
Country
United Kingdom
Harm status
Potential harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Economic or property, Public interest
Harmed
Consumers
Failure modes
Fabricated citation, Training data misuse, Manipulation or dark pattern, Agent misaction, Unsafe recommendation
Consequences
None recorded
AI systems
ChatGPT

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

The Guardian
'Poisoned' AI: the ChatGPT shopping scams that lead to fake websites
7 June 2026
Cybernews
AI chatbot searches leading shoppers to scam websites, consumer groups warn
8 June 2026
Retail Gazette
Scam warning as fake retailer websites appearing in ChatGPT search results
8 June 2026
Futurism
ChatGPT Caught Recommending "Products" That Are Just Scams That Steal Your Credit Card Info
9 June 2026
Mirror
ChatGPT being 'poisoned' into showing scam websites that could drain your bank
9 June 2026
TechSpot
PSA: Scammers are getting their fake shopping sites into ChatGPT's results
9 June 2026
timesnownews.com
ChatGPT recommends scam shopping sites after data poisoning