> The point of this is that you can use the credentials on your phone to prove that you are an adult to a website using zero-knowledge proofs to avoid disclosing your identity to anybody.

No it isn't.

Literally that is not the scope document, and such a solution would not be permitted by the EU as compliant with the legislation.

The app isn't zero knowledge. A prototype workflow has been designed for a one way transfer to sites that is zero knowledge, but it doesn't actually deliver zero knowledge because it you have to verify your age with an external provider to get the credential (which is not zero knowledge), the app has to be secured with either Apple or Google's attestation services (which are not zero knowledge), and the site has to be able to check with the original external provider that the credential hasn't been revoked (which is in no way zero knowledge).

▲

rcxdude 2 days ago | parent | next [-]

The goal would be that neither the verification service nor the service you are verifying with can link the connection: the verification service can't tell which service you are connecting to, and the service you are verifying your age to can't determine your ID. The first two issues you mention don't necessarily seem to kill that (though I agree they are both suboptimal: once you are verified you should be able to generate your own verification keys without connecting to the verification service, and any requirement for attestation is just an unncessary restriction), though the revocation check does seem like it might be a problem.

The issue is that a lot of these services wave around a lot of words that _might_ mean that they are reasonably private, but it's damn hard to actually detemine if it is actually working like that in practice (the eIDAS standard seems to suggest the ZKP stuff is entirely optional, for example).

▲

JimDabell 3 days ago | parent | prev [-]

Zero knowledge proofs are when the prover can prove the statement is true to the verifier without disclosing more information beyond the statement. It doesn’t mean the prover cannot talk to other systems to produce the statement.

▲

Mindwipe 3 days ago | parent [-]

That only works in the context of when the sender isn't the adversary, which isn't the case in an age verification system - it very much does treat the sender as the enemy and untrusted. And again, the revocation chain on the backend is not zero proof.

	▲	rcxdude 2 days ago \| parent [-]
		That's only an issue if getting the proof involves somehow identifying the service you are sending it to. If it's a generic 'send me a proof' it's not necessarily a problem, though of course it would be better if you could just generate your own proof.