Comment by AnthonyMouse

It seems worth thinking past step one if you intend to do something. Even if you had some reliable way to know someone's age, what are you going to do with it in the context of information availability? The proposal is building a privacy-invasive age-leaking system (do you actually want adversarial/malicious services knowing when someone is a vulnerable kid?). There is no point in doing that if the "good thing" it's supposed to enable is actually a hopeless omnishambles.

Meanwhile we don't have any sound technical means of verifying age over the internet. The "use government ID" approaches are among the least effective because you have no good way to tell if the person behind the screen is the person on the ID.