Study Finds LLM Users Have Weaker Understanding After Research

Researchers at the University of Pennsylvania's Wharton School found that people who used large language models to research topics demonstrated weaker understanding and produced less original insights compared to those using Google searches. The study, involving more than 4,500 participants across four experiments, showed LLM users spent less time researching, exerted less effort, and wrote shorter, less detailed responses. In the first experiment, over 1,100 participants researched vegetable gardening using either Google or ChatGPT. Google users wrote longer responses with more unique phrasing and factual references. A second experiment with nearly 2,000 participants presented identical gardening information either as an AI summary or across mock webpages, with Google users again engaging more deeply and retaining more information.

Read more of this story at Slashdot.