Senate Bill 26-051 reflects that pattern. The bill does not directly regulate individual websites that publish adult or otherwise restricted content. Instead, it shifts responsibility to operating system providers and app distribution infrastructure.
Under the bill, an operating system provider would be required to collect a user’s date of birth or age information when an account is established. The provider would then generate an age bracket signal and make that signal available to developers through an application programming interface when an app is downloaded or accessed through a covered application store.
App developers, in turn, would be required to request and use that age bracket signal.
Rather than mandating that every website perform its own age verification check, the bill attempts to embed age attestation within the operating system account layer and have that classification flow through app store ecosystems.
The measure represents the latest iteration in a series of Colorado efforts that have struggled to balance child safety, privacy, feasibility and constitutional limits.
I feel like #1 and #2 are problems whether its client side or server side. As for #3 I would lean in the direction of there being a one-time check with no persistent knowledge. Like when you flash your ID to the bartender to order a drink. A client app that scans the ID and returns the answer to the requestor.
But I don't think there is any way to reliably implement this sort of thing. I think it should really just be left to parental control and monitoring.
I think part of the problem is there shouldn't be a server-side to this. Because that's opening the door to all kinds of intrusive data-collection to determine age, even if they claim it should be done "minimally." Define "minimal." That seems to fly in the face of "clear and convincing information that a user’s age is different than the age indicated by an age signal" which is a direct quote from the Bill.
And as for number 3, I don't see how no persistent knowledge could work. If the client app has read the data ("scanned the ID") that means the client-app can now store that data anywhere the client-app has write access.
Further, it's not like in real life when the bartender can scan the person up and down, look at the ID and make the assessment that McLovin is clearly underage.
If it's open source it can be verified that it's not storing the data.
And I 100% agree that software scanning an ID is an overall bad way to verify. With a CC# validation at least that shows up on my statement, but if my kid is sneaky enough to get mine out of my wallet I have no way of knowing.