They're lying about the rule itself to justify more lies

They know they can't reliably distinguish AI from non-AI by the individual piece of content, and they're pretending otherwise so they can silence actual humans and let chat bots propagate while labeling them as human or AI