My own example
Does AI really get things wrong, even when it sounds sure?
Yes. When I was building the courses for my Academy, I had AI write the first drafts overnight. Four courses were written while I slept. Writing them myself would have taken me about a month.
But every morning I read every word. And every so often it had stated something that simply wasn’t true, or put words in my mouth that I’d never said. Not often, but often enough that if I’d published without checking, my name would have been on things I didn’t believe.
So I don’t stop using it. I give it better information, clearer rules, and I read before anything goes anywhere.