Assuming the company running the service and doing the verification is acting in good faith (big leap here, I get it) couldn’t you verify an identity, store a piece of static information about that person (DL, SSN even tho that sucks) in a hash so that no one else could use that identity to create an account and then issue an account ID with no link to identity marker?

This would allow you to verify users, prevent people from using an ID that was already used, prevent you from being able to link an account to an identity, and prevent you from being able to easily return a list of everyone identified on the service. Best you could do is respond to an individual query on whether that person has verified with your service.

I think it could work technically, but I agree that in practice the US would use its power to make you conduct surveillance without alerting customers, or maybe enact some KYC type requirements for internet usage. This would likely be a first or skipped step on the way to that.