A study finds that people who use “large language models” to research topics had a weaker understanding of those topics ...