AI chatbots are recommending fake retail websites, study finds
Reported by letsdatascience, The Verge, Gadget Review and 1 more outlets4 sources ↓
Podqi tested ChatGPT, Gemini and Perplexity on shopping questions about 151 brands and found risky retail sites in their answers. Across 6,795 answers, Perplexity surfaced them most often and Gemini least often.
Key points
- Perplexity surfaced risky sites in 15% of answers, ChatGPT in 8% and Gemini in 4%
- The tests generated 6,795 answers about 151 brands during two September testing periods
- Of answers containing a risky site, 57% recommended it as a place to shop and 9% warned shoppers away
Sources
- letsdatascienceRead ↗Podqi study finds AI chatbots recommend fake retail websites
- The VergeRead ↗AI chatbots are recommending fake retail websites, study finds
- Gadget ReviewRead ↗AI Chatbots Are Recommending Fake Shopping Websites to Users, New Study Finds
- FuturismRead ↗AI Chatbots Won’t Stop Recommending Scam Websites to People Trying to Shop
Details
Podqi asked the chatbots where to buy brands, which sites carried them and whether they had regional or international storefronts. The firm classified impersonators, counterfeit sellers and certain abandoned or inactive domains as risky. Gadget Review reports that the tests took place in two periods in September.
Futurism reports that 57% of answers containing a risky site recommended it as a place to shop, while 34% did so without comment and 9% warned shoppers away. ChatGPT warned about a risky site in 24% of the instances in which it surfaced one, according to Futurism.
The results show exposure to potentially risky websites, not confirmed fraud or how many shoppers were harmed.
Written from coverage as of
Images
- The Verge
- Gadget Review
- Futurism