4 ms·

Andrew Appel (CS Prof. at Princeton) also gave a very good talk earlier this year about the problems of electronic voting systems. https://www.youtube.com/wat

by pdkl95 10y ago

Andrew Appel (CS Prof. at Princeton) also gave a very good talk earlier this year about the problems of electronic voting systems. https://www.youtube.com/watch?v=abQCqIbBBeM https://www.youtube.com/watch?v=abQCqIbBBeM edit: Also, Ron Rivest was recently on Numberphile, where he talked about using homomorphic encryption to enable end-to-end verifiable voting. https://www.youtube.com/watch?v=BYRTvoZ3Rho https://www.youtube.com/watch?v=BYRTvoZ3Rho

vidarh 10y ago
Voting is explicitly designed not to be end to end verifiable, because if the voter can verify that their vote was recorded correctly, so can someone who wants to buy it or coerce the voter. We instead take extensive precautions with paper votes to physically protect the integrity of ballot boxes and the counting process explicitly to avoid having to have a system where votes can be verified, because there's been plenty of experience with what can happen if they can be. The problem with most electronic voting is that it throws away this verification of integrity of the voting record.
- pdkl95 10y ago
  In the video, Ron Rivest specifically addresses how to allow both an anonymous vote with (some) mutually exclusive end-to-end verification. I completely agree that electronic voting is currently a terrible idea. The other talk I linked to by Andrew Appel strongly encourages using only paper votes and statistically verifying them by manually recounting a random sample of voting locations. However, I'm willing to listen to Ron (the "R" in "RSA") Rivest's ideas about using encryption in voting. I think the idea needs more work, but it's interesting approach that does attempt to address both anonymity and verifiability.
  - marten-de-vries 10y ago
    It's worth noting that using encryption techniques makes the functioning of a voting system opaque to 99% of the population. In contrast, everyone can understand how voting using paper works. It's no doubt possible to make digital voting systems with some awesome properties, but in the end, are the improvements they give worth this loss of understanding (which might come with a decreased trust in the system)?
- sparky_z 10y ago
  > because there's been plenty of experience with what can happen if they can be. Such as? Was there ever a time in the history of modern democracies where "vote receipts" were actually used and caused problems? I was under the impression that they've never been used out of preemptive fear.
  - dmurray 10y ago
    Voting in the United Kingdom was public until 1872[0]. After the change to secret ballots, a lot more worker- and tenant- friendly representatives were elected in Britain and Ireland, and it is generally credited with enfranchising the Irish people. [0]https://en.m.wikipedia.org/wiki/Ballot_Act_1872 https://en.m.wikipedia.org/wiki/Ballot_Act_1872
- heinrich5991 10y ago
  Suppose you were told a random number between 1 and (number of vote options) when you are in the booth. Later, your choice plus that random number is published, then you (but no one else) can verify the result, because you can trivially tell them any other number that modifies the result to something you want.
  - aninhumer 10y ago
    Someone could ask you to disclose your number before the election results are published.
    - heinrich5991 10y ago
      And you can immediately calculate the number that you need to tell to fake any choice you want.
      - aninhumer 10y ago
        I don't think this works: Suppose there's a method for calculating a fake choice token from a true choice token. This means that applying it to a fake choice must also produce corresponding fake choice tokens, otherwise a fake choice could be identified in this way. And by induction, any number of applications of the function must also reach a matching choice. (This isn't a complete proof, but chasing up the remaining loose ends makes me think such a system is impossible, or at very least extremely complicated, and almost certainly impractical for non-mathematicians.)
        heinrich5991 10y ago
        Say there are two vote options, 0 and 1. We use addition modulo 2, i.e. XOR. I vote for candidate 0. The voting machine tells me the random number 1 (out of 0, 1). When the voting results are published, I can check that next to my name it has 0^1 = 1. When someone asks me what my random number is, I can either answer 0, to pretend I voted for 0(random number)^1(public list) = 1, or 1, to say that I voted for 1(random number)^1(public list) = 0.

by pdkl95 10y ago

vidarh 10y ago

Voting is explicitly designed not to be end to end verifiable, because if the voter can verify that their vote was recorded correctly, so can someone who wants to buy it or coerce the voter. We instead take extensive precautions with paper votes to physically protect the integrity of ballot boxes and the counting process explicitly to avoid having to have a system where votes can be verified, because there's been plenty of experience with what can happen if they can be. The problem with most electronic voting is that it throws away this verification of integrity of the voting record.

pdkl95 10y ago
In the video, Ron Rivest specifically addresses how to allow both an anonymous vote with (some) mutually exclusive end-to-end verification. I completely agree that electronic voting is currently a terrible idea. The other talk I linked to by Andrew Appel strongly encourages using only paper votes and statistically verifying them by manually recounting a random sample of voting locations. However, I'm willing to listen to Ron (the "R" in "RSA") Rivest's ideas about using encryption in voting. I think the idea needs more work, but it's interesting approach that does attempt to address both anonymity and verifiability.
- marten-de-vries 10y ago
  It's worth noting that using encryption techniques makes the functioning of a voting system opaque to 99% of the population. In contrast, everyone can understand how voting using paper works. It's no doubt possible to make digital voting systems with some awesome properties, but in the end, are the improvements they give worth this loss of understanding (which might come with a decreased trust in the system)?
sparky_z 10y ago
> because there's been plenty of experience with what can happen if they can be. Such as? Was there ever a time in the history of modern democracies where "vote receipts" were actually used and caused problems? I was under the impression that they've never been used out of preemptive fear.
- dmurray 10y ago
  Voting in the United Kingdom was public until 1872[0]. After the change to secret ballots, a lot more worker- and tenant- friendly representatives were elected in Britain and Ireland, and it is generally credited with enfranchising the Irish people. [0]https://en.m.wikipedia.org/wiki/Ballot_Act_1872 https://en.m.wikipedia.org/wiki/Ballot_Act_1872
heinrich5991 10y ago
Suppose you were told a random number between 1 and (number of vote options) when you are in the booth. Later, your choice plus that random number is published, then you (but no one else) can verify the result, because you can trivially tell them any other number that modifies the result to something you want.
- aninhumer 10y ago
  Someone could ask you to disclose your number before the election results are published.
  - heinrich5991 10y ago
    And you can immediately calculate the number that you need to tell to fake any choice you want.
    - aninhumer 10y ago
      I don't think this works: Suppose there's a method for calculating a fake choice token from a true choice token. This means that applying it to a fake choice must also produce corresponding fake choice tokens, otherwise a fake choice could be identified in this way. And by induction, any number of applications of the function must also reach a matching choice. (This isn't a complete proof, but chasing up the remaining loose ends makes me think such a system is impossible, or at very least extremely complicated, and almost certainly impractical for non-mathematicians.)
      - heinrich5991 10y ago
        Say there are two vote options, 0 and 1. We use addition modulo 2, i.e. XOR. I vote for candidate 0. The voting machine tells me the random number 1 (out of 0, 1). When the voting results are published, I can check that next to my name it has 0^1 = 1. When someone asks me what my random number is, I can either answer 0, to pretend I voted for 0(random number)^1(public list) = 1, or 1, to say that I voted for 1(random number)^1(public list) = 0.