> Theoretically, if you did trust your government,
Trust not to misuse my data intentionally?
Trust to not lose my data due to incompetence?
Trust to not subcontract this service to some spying company like Google?
> couldn't zero-knowledge proofs be used to allow such a system?
They could, but will they really be zero knowledge or there will be some intermediary leftovers that aren't zero knowledge? See trust in competence above.