California lawmakers unanimously pass Linux exemption from age-verification law

First, leaking a device's content filtering settings is not the same as leaking a user's age bracket. For example, if the server identifies a web browser that isn't loading tags annotated with "sexual-content" that might indicate that the user is a child but could equally well indicate that they're a corporate office worker, religious, or anyone else who prefers not to see such content at the moment.

Second, if the standard defined multiple semantic tags or levels (e.g. violence, extreme violence, substance use, unknown, etc) then this gives more granular control to device administrators who may care more about some types of content than others.

Third, web browsers wouldn't even have to leak the content filtering settings. For example, perhaps an administrator could configure the web browser to render content overlaid with a semi-opaque blur, in which case the server would not know.

Fourth, and this is more philosophical, asking websites to provide more information to the device so that an admin can do their own filtering leaves the choices to them. Whereas if the only way a device admin can filter content is to submit the age of the device's user to a third party so the third party can decide how to pre-filter the content it sends back, that gives an uncomfortable amount of control to platforms and regulators.