Why It Matters
How Scammers Infiltrate AI Search
The Numbers
- 55% of adults use AI search weekly, according to a 2026 Pew Research Center survey (Pew Research Center).
- 2.4 million fraud complaints were filed with the FTC in 2025; online shopping scams were the second-most reported category (FTC Consumer Sentinel Network Data Book 2025).
- At least two cloned retail sites surfaced directly inside ChatGPT results, according to Ask Silver’s investigation (Ask Silver).
- Up to 80% discounts were advertised on the fraudulent pages—a classic bait tactic—while checkout forms collected bank details.
- 42% increase in purchase scams was reported by the UK’s National Fraud Intelligence Bureau in 2025/26, reflecting a sharp rise in online shopping fraud (Action Fraud / NFIB).
“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.”
— Louise Baxter, Head of the Scams Team, National Trading Standards (National Trading Standards)
What Comes Next
What This Means for You
claiming and verifying your business BizScoreAI’s Get Listed tool AI Contactability scan 88% of AI Mode users accept results as-is AI Search Behavior Study German Court AI Overviews Ruling Feedsta.aiThe Bigger Picture
AI search convenience can silently hand scammers a brand’s credibility; if your business isn’t verifiably real, a clone will fill the gap.
Frequently Asked Questions
How did fake Russell & Bromley sites end up in ChatGPT results?
Scammers built a convincing clone site with a lookalike domain and published it online. When ChatGPT’s underlying model was trained or when it retrieved live data, it treated that site as a legitimate source. Because the real Russell & Bromley brand no longer operates a standalone website after being absorbed by Next, there was no strong official signal to outrank the fake domain, allowing the scam page to surface in shopping queries.
Can AI search platforms prevent fraudulent links from appearing?
Currently, they rely on after-the-fact reporting and manual removal. ChatGPT removed the flagged domains once notified, but proactive prevention requires source verification at crawl time. AI companies are exploring credibility scoring and domain verification partnerships, but none yet deploy a foolproof gate. Until that changes, fraudulent sites will occasionally appear in AI-generated shopping results.
What should a business owner do to stop scammers from cloning their brand in AI search?
The first defense is a consistent, verified online footprint. Claim your business on all major directories and data aggregators, ensure your website uses HTTPS and structured data, and regularly search for your brand name in AI-powered tools like ChatGPT or Perplexity. Services like BizScoreAI’s Get Listed tool and AI Contactability scan help identify gaps that scammers could exploit before they become a problem.
Is it only ChatGPT that shows fake shopping sites, or are other AI searches affected?
The underlying mechanism—training on or retrieving from the open web—applies to any AI search tool. While this incident was documented in ChatGPT, similar risks exist in Google AI Overviews, Perplexity, and models with web browsing capabilities. Any AI that compiles links from crawled pages can inadvertently include malicious sites.
How can consumers spot an AI-recommended site that might be a scam?
Check the domain carefully for misspellings or added words, look for the padlock icon and a legitimate SSL certificate, and avoid sites that demand payment by bank transfer instead of standard card processors. When an AI reply includes a link, visit the retailer’s homepage via a known URL rather than clicking the embedded link. Suspicious sites can be reported through ChatGPT’s reporting form or to national fraud bodies.
Is there a link between brand closures and AI search scams?
Yes—scammers often target brands that have closed or been acquired because the official web presence disappears, leaving a vacuum in search indexes. In this case, Russell & Bromley’s administration meant no official site existed, so AI models latched onto whatever they could find. That pattern makes post-merger brand integrity and redirects critical for preventing impersonation.
Sources
Is your business visible to AI assistants?
Check Your Score →
Run a free scan to see your AI Visibility Score, SEO rating, and local citation accuracy.