Chatbot check: ChatGPT, Gemini and Perplexity may lead to fake stores

According to the results of a new analysis, services such as ChatGPT, Gemini and Perplexity may direct users to fraudulent resources or counterfeit copies of well-known online stores.

Risk levels across different platforms

Experts checked chatbot recommendations for 151 brands. Suspicious links were found in the responses of all three services studied. Perplexity led in the number of problematic recommendations: dangerous links appeared in 15% of cases.

The figure for ChatGPT was lower — 8%, while Gemini demonstrated the best result among those checked, with the share of suspicious links at 4%.

Recommendations to dangerous sites

Particularly alarming is the fact that in 57% of cases where a dangerous site appeared in the response, the artificial intelligence simultaneously contained an encouragement to use it for making purchases.