Attempting to prevent websites from learning about their audience is difficult. For example, Lego has a "play zone" and it has other pages geared towards adults. You get an immediate popup asking if you want to go to the play zone or not. If a user clicks on play zone, that's a good signal (but not guaranteed) that they're probably a child.

Preventing this would be difficult. If there are any behavioral differences between children and adults, that's a signal. And it would mean websites couldn't build experiences tailored to their audiences. Do we really want to try to suppress all behavioral cues that can be used to distinguish children from adults? Is it worth it?

A reasonable compromise is that websites and apps should be able to learn about devices, not people. It should be trivial to figure out whether a visit is from a child-locked device or not, no account needed. Much like clicking on "kids zone," it doesn't mean you know for sure who's really using the device. For example, it could be an parent who turned the child lock on to see what happens, because parents should know these things.

It's another bit that could be used distinguish users, but giving away zero bits of information isn't practical.