AI tools can mislead with errors known as hallucinations, ranging from 22% to...
https://wiki-planet.win/index.php/Should_a_Family_Office_Ban_ChatGPT_for_Client_Work%3F
AI tools can mislead with errors known as hallucinations, ranging from 22% to 94% in some cases. For family offices managing complex wealth, this means nearly half of AI outputs may be unreliable, with up to 20% containing material mistakes