Affiliate links on Android Authority may earn us a commission. Learn more.
ChatGPT was caught recommending fake scam stores
- ChatGPT shopping results have been linked to fake stores, exposing some users to financial fraud and stolen payment details.
- Scammers exploited the disappearance of Russell & Bromley as an independent retailer by creating convincing copycat websites.
- Researchers suspect “AI poisoning” is helping fraudulent pages infiltrate large language models and appear in recommendations.
AI shopping assistants promise to make buying products easier, but a growing wave of fraud shows they can sometimes bring whole new risks.
The Guardian has reported that there have been instances where ChatGPT has led users to convincing fake retail websites and even suggested products that were never real. According to findings shared by scam-detection service Ask Silver, cloned online stores have started to appear in ChatGPT-generated shopping results. Fraudsters operated websites that looked authentic but actually featured links to popular brands such as Russell & Bromley or Dunelm for users searching for them. Those who placed orders lost money but also had their payment details exposed.
One reason for the confusion is that Russell & Bromley ceased to exist as a standalone retailer after going into administration in January 2026 and being swallowed up by Next. Scammers jumped in to fill the void with lookalike pages, as shoppers still looked for the brand’s original website.
Don’t want to miss the best from Android Authority?
- Set us as a favorite source in Google Discover to never miss our latest exclusive reports, expert analysis, and much more.
- You can also set us as a preferred source in Google Search by clicking the button below.
These attacks may use what researchers think is “AI poisoning,” a technique where bad actors inundate the web with fake information and cloned pages that are eventually absorbed by large language models. It unwittingly promotes scams by promoting fake sources as if they were real.
Consumer protection experts say people shouldn’t assume a recommendation can be trusted just because it’s coming from an AI chatbot. Criminals are quick to adapt to new technology and will use any channel that gives them access to potential victims, said Louise Baxter, head of the scams team at National Trading Standards.
The warning signs are still the same. Massive discounts, strange website addresses, bad contact information, and requests for bank transfers should all be suspect. Experts recommend going directly to the sites of the official retailers instead of relying on AI-generated links.
Some of the flagged sites have been taken down by OpenAI after being reported, but the bigger problem remains. With AI tools playing an even bigger role in how people search and shop online, companies will require stronger safeguards to prevent scammers from turning helpful recommendations into costly mistakes.
Thank you for being part of our community. Read our Comment Policy before posting.