pallest·3 anni fa·discussOne solution is placing the ChatGPT response in another model that determines the accuracy of the information and revises where necessary