Yes, I will help you solve the problem.
You say that Person A used genAI to generate an email, and Person B used genAI to summarise that email, and it seems that Person A thinks that they asked Person B to do X, but in fact their email said Y, and Person B thought that Person A had asked them do Z but in fact their email said Y.
Yes, I wonder where the problem might be. Hard to tell, isn't it.
@neil I still feel that two groups will make bank off of LLMs, the legal field and medical psychology..... — Vi (basing this on the best way to make money off of the stockmarket being selling traders something they need to be better at trading)