I just saw this, and... welp.

Reading this, there's another thing I can't help but notice that I haven't seen touched on. Namely, part of the process that the AVPA spokesperson was describing involves making sure that subsequent re-checks are done by the same user who completed the initial check:

How often you need to prove it is still the same user who did the check is a matter for the services themselves and their regulators. Some low risk uses might only check every three months – higher risk situations might double check it is still you each time you make a purchase.

This implies that the service would have to hold enough (presumably biometric) data about you to make this determination - which by definition means it's personally identifiable information, even if they can't get a name from it.

It doesn't take much to imagine a scenario where law enforcement could give such a service a high-resolution photo of the face of someone they're interested in and get back a record of which sites that person has been validated on, and when.

[edit: And while I do realise this is one of the least horrifying things about this whole thing, it's... still pretty horrifying, which says a lot about everything else about this.]