ChatGPT is suggesting fraudulent websites that could take your credit card information.

      The chatbot is revealing counterfeit versions of extinct retail brands, with scammers intentionally creating websites to manipulate its suggestions.

      Scammers have discovered a new method to reach consumers: utilizing ChatGPT for their marketing purposes. As reported by The Guardian, the scam verification service Ask Silver has identified that OpenAI's chatbot is directing users to fraudulent retail sites designed to collect payment information from unsuspecting shoppers. These sites replicate legitimate storefronts and employ official-looking URLs, making them hard to distinguish without careful examination.

      Inactive brands are a primary target

      It seems that fraudsters are intentionally focusing on brands that have recently gone out of business or been taken over, creating a disconnect between consumer interest and a legitimate online presence.

      Anna Jones from Ask Silver informed The Guardian about Russell & Bromley, the British shoe retailer that entered administration in January 2026, as a notable case. With no official website after Next absorbed the brand, scammers constructed a convincing replica and optimized it to appear in ChatGPT’s results. Consumers inquiring about Russell & Bromley products by name are directed to this fraudulent site.

      Jones mentioned that one possible reason for this issue is that malicious actors may have introduced content into ChatGPT’s training data to promote these counterfeit pages, a practice sometimes referred to as data poisoning.

      OpenAI seems to have revised its search results for inquiries related to Russell & Bromley, now displaying a warning that states, “Several websites currently advertising Russell & Bromley products at a significant discount appear suspicious and may not be official retailers. Recent reports have indicated the emergence of fake Russell & Bromley stores in AI-driven search results.”

      A growing concern as AI becomes more involved in shopping

      Research from last year indicated that ChatGPT already has difficulty providing consistent and reliable product recommendations, but redirecting users to fraudulent websites is a more critical issue. The situation is likely to worsen as AI plays an increasingly active role in the purchasing process.

      Louise Baxter from the UK’s National Trading Standards told The Guardian that scammers quickly adapt to advancements in technology. With the emergence of agent-driven AI shopping still in its infancy, there is a limited timeframe to address this vulnerability.