A study by Product.ai found that major large language models had high error rates when used for online shopping. The company tested free and paid versions of ChatGPT, Claude, Gemini, and Perplexity with 220 shopping questions, capturing 8,794 responses. Problems included giving wrong prices, incorrect product specifications, and not finding the latest models. The study suggests AI is not yet ready to take over holiday shopping.
Related Articles
Don't miss out on breaking stories and in-depth articles.