End-to-End Verifiable Postal Voting (Transcript of Discussion)

Abstract

Conclusion: So the argument is supposed to be that this process of randomly choosing one link or the other is supposed to give you confidence that the set of pieces of paper on the far left-hand side that the voter has actually looked at matches the set of encrypted things on the right-hand side that the voter has signed with their public key. But you didn’t actually get to learn for any individual vote which signature matches which plaintext.

Michael Roe: So the human-readable one isn’t identifiable as to which vote it is?

Reply: Correct. It’s just a completely ordinary vote. In Australia, this is an ordinary piece of paper with a vote on it and without a unique ID. More questions? The obvious question is, what’s the process for choosing this random bit in a way that makes anybody believe that this process is being done correctly? Because clearly if you can manipulate the bit-choosing process, then you can also manipulate the vote, one way or another, even without being able to forge signatures.