Cloned retail sites seep into AI shopping results

Prompt fraud fears as consumers shift to AI search for product picks

AI
Google search engine
  • Investigators say scammers are exploiting retail transitions, including store closures and site changes, to capture brand searches.
  • Researchers warn such “data poisoning” campaigns are escalating as criminals attempt to influence how AI systems rank retailers, products, and deals.

Cloned retail websites are appearing in AI-generated shopping recommendations, raising questions about how effectively AI search tools are filtering fraud as more consumers rely on them for product picks.

UK scam-checking service Ask Silver identified fake sites impersonating British retailers Russell & Bromley and Dunelm being surfaced as sources in ChatGPT responses. In tests asking for popular Russell & Bromley bags and purses, ChatGPT provided product suggestions, prices, and links—some of which routed users to fraudulent lookalike domains.

Investigators say scammers are exploiting retail transitions, including store closures and site changes, to capture brand searches.

Data poisoning campaigns

Ask Silver suggested confusion following Russell & Bromley’s closure and subsequent acquisition by Next earlier this year may have created openings for clones that mimic branding, imagery, and product descriptions.

Some fake sites dangle discounts of up to 80 per cent—a classic red flag—and use domains such as therussellbromleyofficial, russellandbromleylondon, russellbromleyonlineuk, and russell-and-bromley.

Advertisment

Ask Silver co-founder Anna Jones said the underlying large language model may have been “poisoned,” referring to tactics where malicious content is injected into web sources—via cloned pages, synthetic reviews, and automated forum posts—to manipulate what AI systems crawl and recommend.

Researchers warn such “data poisoning” campaigns are escalating as criminals attempt to influence how AI systems rank retailers, products, and deals.

Louise Baxter, head of scams at National Trading Standards, cautioned consumers not to assume legitimacy simply because a site appears in an AI-generated answer, noting criminals are adapting quickly. OpenAI told The Guardian it removed the fraudulent sites from its search index.

Consumer advice:

  • Verify URLs carefully and navigate to official retailer sites directly.
  • Treat steep discounts (e.g., up to 80%) as high-risk indicators.
  • Look for secure payment methods and check independent contact details and returns policies before purchasing.

Discover more from TechChannel News

Subscribe to get the latest posts sent to your email.