This has, again, already been answered. We are not intending to use AI to detect it. AI itself writes in very distinct ways, far different from how you do. It's nothing to do with standards when the difference in standards between human input and AI are quite obvious. If it's too hard to tell, we do not act. The initial explanation of the new rule details this, as have several people in the thread including demonstrations of what is and isn't acceptable.
I urge you to read those responses and otherwise try to relax, Helmchen.