As long as you have some automated AI bot sent all the reports. It’s never fraud, you couldn’t have known it would do that.
It's like having a dangerous dog that usually doesn't bite, but you really cannot know if it will change its mind one day. Do you just let such dog walk the streets without owner supervision?
In other words, a pit bull.
It’s irresponsible to deploy AI if you don’t know what it will do, especially when there are actual stakes.
Maybe we’ll have less AI bullshit then.