Artificial Intelligence has rapidly transformed the way people shop online. From comparing products to finding the best deals, AI-powered tools such as ChatGPT are increasingly becoming digital shopping assistants for millions of consumers. However, a recent investigation has revealed a worrying downside to this convenience: scammers are exploiting AI systems to direct shoppers toward fraudulent websites.
The findings have sparked fresh concerns about the growing sophistication of online scams and the challenges AI platforms face in keeping bad actors out of their search results.
Credits: NPR
Scammers Are Using AI’s Growing Influence
Consumers often trust recommendations generated by AI tools, assuming the information provided is accurate and reliable. Cybercriminals are now leveraging this trust to their advantage.
According to research conducted by Ask Silver, fraudsters are creating cloned websites that closely mimic legitimate retailers. These fake websites are designed to appear authentic, complete with professional layouts, product listings, and promotional offers. Unsuspecting shoppers who discover these sites through AI-generated recommendations may believe they are dealing with a genuine brand.
The tactic becomes particularly effective when a retailer has gone out of business, changed ownership, or merged with another company. In such cases, customers searching for a brand online may struggle to identify the official website, creating an opportunity for scammers to fill the gap with convincing imitations.
The Growing Threat of AI Manipulation
Experts believe that some of these fraudulent websites may be influencing the information that AI systems rely on when generating responses.
Anna Jones from Ask Silver suggested that the large language model powering ChatGPT may have been “poisoned” by misleading online content. In simple terms, bad actors can flood the internet with fake pages, deceptive information, and cloned websites, increasing the chances that AI tools encounter and reference them.
This practice, often referred to as data poisoning, represents a new frontier in online fraud. Rather than targeting consumers directly through emails or advertisements, scammers are attempting to manipulate the information ecosystem itself.
As AI tools become more integrated into everyday decision-making, ensuring the quality and reliability of online data will become increasingly important.
Research Uncovered Fraudulent Retail Websites
The investigation highlighted a real-world example involving British footwear and fashion retailer Russell & Bromley.
When researchers asked ChatGPT a general question about popular Russell & Bromley bags and purses, the AI provided product information and pricing details. However, among the sources surfaced were two fraudulent websites impersonating the retailer.
The fake sites appeared professional and trustworthy at first glance. They also featured massive discounts on luxury items—a common tactic used by scammers to create urgency and attract bargain-hunting consumers.
Such discounts can often be a warning sign. If an offer appears significantly better than what is available elsewhere, shoppers should take extra time to verify the website’s legitimacy before making a purchase.
How Consumers Can Protect Themselves
While AI-powered shopping tools remain useful, experts advise consumers to maintain a healthy level of skepticism.
One of the easiest ways to identify a cloned website is by carefully examining its web address. Legitimate retailers typically use established domains, while fraudulent websites may include extra words such as “official,” “sale,” or “deals” in an attempt to appear genuine.
Consumers should also be cautious of websites that only accept bank transfers or direct payments instead of offering secure payment methods such as credit cards or trusted payment gateways.
Another red flag is unusually steep discounts that seem too good to be true. In many cases, these offers are designed to pressure shoppers into making quick decisions without conducting proper checks.
If financial information has already been shared with a suspicious website, experts recommend contacting the bank immediately and reporting the incident.
Credits: BBC
AI Platforms Respond to the Threat
The retailer that acquired Russell & Bromley earlier this year confirmed that it was aware of the fraudulent websites and had been working to get them removed.
Meanwhile, OpenAI stated that the identified fraudulent websites have been removed from its search index. The company also noted that users can report websites that violate its policies through a dedicated reporting process.
The incident serves as a reminder that while AI can simplify online shopping, it should not replace basic digital safety practices. As scammers continue to evolve their tactics, consumers must remain vigilant and verify information independently before making purchases.
