If only they could develop some kind of machine learning AI tech that could make highly suspicious comments require approval.

I guess they just aren't sophisticated enough to do that... But I guess they must have AI that can identify kids in videos accurately (even 17 1/2 year olds).