We don't need a study to tell us that LLMs always make mistakes. We already knew that. Anyone with sense is not using LLMs because of that.