AI tools can sometimes produce “hallucinations”—incorrect or misleading...

https://online-wiki.win/index.php/Should_a_Family_Office_Ban_ChatGPT_for_Client_Work%3F

AI tools can sometimes produce “hallucinations”—incorrect or misleading information—which is a real concern for family offices managing sensitive data. Studies show hallucination rates can range from 22% up to 94%, with accuracy sometimes around only 50%