These could easily be ferreted out by synthetic agents scouring the internet and posing as kids.
Could you give an example of such an attack that is intractable in the status quo?
Well, currently ads can be targeted using user agents. This is analogous to that sort of attack, although less malicious. There are bad people who would target children if they could, so it does seem to fit with what you're asking. We create the means, and they will exploit it.