Don't get me wrong, I'm all for privacy. But between setting up the birthdate when creating my children's local account on their computers, and having to send a copy of their ID to every platform under the sun, I'd easily chose the former.
I'd even agree to a simple protocol (HTTP X-Over-18 / X-Over-21 headers?) to that.
But even if it happened purely at the OS level, it would be laughably unenforceable at best.
Essentially, I’m confused why the world gave up so quickly on parental controls (not really confused—the alternative provides more surveillance capability).
I can easily imagine a piece of software parents can download, for free, that if installed would basically function akin to a virus on someone's computer - it blacklists much of the internet and is updated and maintained by a company that updates the allowed sites and banned sites regularly. It could not be turned off. If it crashes, is ended by force, it automatically reloads - and any attempts to remove it sends emails or text messages to the owners (the parents) who would know something is up. It could be turned off only by the parent putting in a specific password to disable it, and if they forgot, they would have to phone the company to get it reset.
Any responsible parent would install this on the phones and computers of their kids and it would do everything they need.