Sjors's avatar
Sjors 2 weeks ago
I appreciate that the Dutch government at least open sources their dystopian age verification wallet project. It makes it possible to more accurately pin down where the privacy and censorship problems are. I would encourage others in the EU to look for what their country is up to. First concrete issue found: they use a deterministic hash of the national identification number (SSN-like) as a recovery (and blacklisting) mechanism.

Replies (7)

Sjors's avatar
Sjors 2 weeks ago
I'll give the authors some time to respond, also to make sure I'm not hallucinating. With some help from Claude I already found other potential issues, as hinted in the footnotes, but I want to more thoroughly understand these.
Sjors's avatar
Sjors 2 weeks ago
Fwiw I, yesterday I probably black-pilled Claude with my prompts. Hence the need to await reality checks. Meanwhile I should probably read what actual security researchers have written so far. View quoted note →
Sjors's avatar
Sjors 2 weeks ago
The one-to-one coupling does it, yes. It would still be the case if they follow my suggestion to generate (and store) a random number for each BSN and then hash that. It's not fundamental to the EU directive afaik, it's a problem with this particular architecture. But I suspect that any conceivable architecture that's permitted under the law (which is flexible) is going to have serious issues like this.
Sjors's avatar
Sjors 2 weeks ago
I haven't studied that part of the Directive yet. Recovery could just mean: if you lose your pin, you sign up again, download your digital drivers license and other stuff again. Not much different from losing a physical wallet. It's a UX trade-off, but with serious privacy consequences. The harder (but maybe not impossible) part is blacklisting (by the authorities) and revocation (by the user, after they lose access).