The most dangerous thing about modern AI isn't that it gets things wrong.
It's that it gets things wrong in exactly the same voice it uses when it's right.
I learned this the hard way last week.
https://www.sistnt.com/blog/when-ai-tricks-you-with-swag/