Skip to content
Zenteck
Latest
Analysis

Are You Shopping via AI? Stop Trusting Those Links

ByAlex Silva2 min read0 comments
ai-chatbots-suggest-fake-retail-sites
Photo: ZenteckAI chatbots suggest fake retail sites

Listen to article0:00 / 0:30

The hidden danger in your AI shopping assistant

Using AI to find the best deals or regional brand sites is becoming a standard workflow, but a recent study by IP protection firm Podqi reveals a significant security blind spot. When prompted to find where to buy specific brands, major chatbots frequently point users toward phishing sites, knockoff retailers, or expired domains now owned by malicious actors.

According to the findings, the risk is not uniform across platforms. Perplexity was the most prone to error, surfacing risky URLs in 15% of its responses. ChatGPT followed at 8%, while Google’s Gemini performed best, with a 4% error rate. Crucially, smaller brands are twice as likely to be associated with risky links compared to major global retailers.

Why this happens

AI models are trained to be helpful and conversational, which often overrides the strict verification required for e-commerce. When asked for regional brand websites, the models often hallucinate or guess based on outdated training data. The study noted that chatbots sometimes invent fake regional domains for brands that do not even have them, creating a perfect trap for unsuspecting users.

How to protect your data

Treating AI as a search engine for commerce is a security risk. If you use these tools to research products, follow these steps to stay safe:

  • Verify the URL: Never click a link provided by an AI without checking the domain against the brand's official social media or verified website.
  • Check for warnings: ChatGPT was the only tool in the study that provided warnings about risky sites in 24% of cases. If the AI flags a site, do not ignore it.
  • Avoid direct input: Never enter credit card or personal information on a site discovered solely through an AI chat recommendation.

Verdict: Is it worth the risk?

For research, AI is convenient; for transactions, it is currently dangerous. The convenience of a quick link does not outweigh the risk of credit card theft. Until these models implement real-time verification against reputable domain registries, treat every shopping link provided by an AI as potentially hostile.

Sources

  1. AI chatbots are recommending fake retail websites, study finds — theverge.com

Frequently asked questions

Which AI chatbot is safest for shopping research?
In the Podqi study, Google's Gemini performed best, showing risky sites in only 4% of responses, compared to 15% for Perplexity and 8% for ChatGPT.
Why do AI chatbots suggest fake websites?
Chatbots often rely on outdated training data or hallucinate regional domains for brands that don't exist, failing to distinguish between legitimate retailers and malicious imposters.

AI Briefing in 3 Minutes

Get what actually matters in AI & tech delivered free to your inbox. Zero spam, 1-click unsubscribe.

By subscribing, you agree to our Privacy Policy.