We've tried this before. In 1997. https://en.wikipedia.org/wiki/Platform_for_Internet_Content_Selection

The push for this form of age verification is to "close the loop" of strong identity between a physical person, and online content access. And unfortunately, all the relevant parties are aligned on this one:

• Site owners want to prevent non-people AI from accessing their content
• Authoritarian leaning governments want to track and control who gets access to what, and who can even make content available in the first place
• Ad Companies want to make sure ads are being served to real people
• Operating system vendors can kill open source by ensuring this is mandated at the OS level, in a non-bypassable way, so signed/TPM/closed source nonsense.
• Parents who don't understand the technology enabling this but believe that this 'protects the kids' will also support it.

Content should be able to send a request asking the OS if you’re old enough to view

You've got it backwards. This is letting them frame the narrative. There's no need to tell anyone else anything about yourself when requesting content. The local device should be where any content filtering happens.

If you're in a kid mode, it should block the incoming content, not rely on a third party to give a shit about blocking itself.

No chance for "whoops we fascismed" that way

The reason a lot of us take issue with the age-verificarion steps is that while the stated intent is saftey, we are prtty sure none of those guard rails will be implemented. Noone will enforce the fines of telemtry harvesting and fines not high enough are just the cost of doing business.

Noone trusts the entities pushing for these changes to be good stewards of these additional data points and most tech experts are skeptical that they will be effective to begin with. The benifits are negligable and the erosion of user privacy (as pitiful as it is with everyones personal wiretaps) is unacceptable.

Parents and tech companies, if they want features like this need to go back to the drawing board and try again, use the parental control features already embedded in devices and stop trying to shove Orwellian shit into every new requirement.

Aside from the problems stated in the other replies, another problem are the differing standards of what kids should or shouldn't access.

There are some websites some parents would be fine with their kids accessing that other parents could consider obscene. Given that much of the argument is giving back control to parents, the definition of 'obscene content' shouldn't be decided for them. This is especially true given that if conservatives had their way, all LGBTQ content would also be age restricted.

Why not other way around? Website send which age appropriate and OS is set by admin to not display if age not match?

Edit: Or make standard list of "content tag" which classify content for parent to deckide? And fine site out the ass if not send correct tag. Decision 100% at parent hand no telemetry possible, except maybe if keep loading site but can only confirm no tag blocked.