There is no such thing in practice.
Anything with zero knowledge is never going to be considered robust enough by a government. Zero knowledge protocols really have no functional revocation mechanism.
(Without accepting the premise that it should be acceptable to have to provide any kind of proof...)
> Zero knowledge protocols really have no functional revocation mechanism.
None would be needed, you (sadly) only age in one direction, so valid proof would never become invalid proof.
expiry
The EU has been working on a zero knowledge system as part of the EU Digital Identity Wallet project for a few years now. It is currently undergoing large scale field tests in several countries with expected release late this year. All member states are required to provide at least one free secure interoperable implementation to their citizens, and regulated industries such as banks and telecoms, are required to accept it. If a member state passes a law requiring age verification on social media it must include the EU Digital Identity Wallet as one of the verification methods the site must support.
What was that about no government would consider zero knowledge to be robust enough?