AI chatbots are recommending fake retail websites, study finds

The convenience of asking an AI assistant where to buy a specific product is undeniable. However, a concerning new study suggests that this convenience comes with a significant safety risk. Shoppers using popular chatbots might be steered toward imposters and risky websites instead of legitimate retailers.

According to reporting by The Verge, the research was conducted by IP protection firm Podqi. They tested three leading AI models: ChatGPT, Gemini, and Perplexity. The goal was to see how these systems handle commercial queries that could lead users to fraudulent e-commerce platforms.

The testing methodology was straightforward but revealing. Podqi posed several shopping-related questions involving 151 different brands. These ranged from small companies to large global corporations. The questions included standard queries like where to find a brand online or if they ship internationally.

The tests ran over two periods in September. This timing is notable as it captures the AI landscape during a peak shopping season. The results showed that these powerful language models are not always reliable guides for commercial transactions. They can inadvertently promote unsafe digital storefronts.

This finding underscores a broader issue in the current state of generative AI. While these models excel at creative writing and coding, they struggle with real-world verification. They often prioritize plausible-sounding answers over factual accuracy. This is particularly dangerous in the retail sector where financial loss is a direct consequence of bad advice.

For entrepreneurs and professionals, this is a wake-up call. We cannot blindly trust AI outputs for high-stakes decisions. The integration of AI into daily workflows requires a layer of human verification. This is especially true when money or sensitive data is involved.

What this means for you is that you must treat AI recommendations as starting points, not final answers. Always verify the URL and the legitimacy of a website before entering any payment information. To stay safe, try using this prompt with your AI assistant to vet a site: "Analyze the domain [insert URL] for common signs of phishing or fraud, such as recent registration dates or mismatched SSL certificates, and list three red flags to watch for."