Scientists question electronic voting

[Ed Gerck]

Anton Stiglic wrote:

> An extortionist could provide their own camera device to the voter, which
> has
> a built in clock that timestamps the photos and does some watermarking, or
> something like that, which could complicate the counter-measures. But this
> problem already exists with current non-electronic voting scheme.
> It depends on the value attributed to a vote (would an extortionist be
> willing to provide these custom devices?).

This is not possible for current paper ballots, for several reasons. For
example, if you take a picture of your punch card as a proof of how you
voted, what is to prevent you -- after the picture is taken -- to punch
another hole for the same race and invalidate your vote? Or, to ask the
clerk for a second ballot, saying that you punched the wrong hole,
and vote for another candidate?  The same happens for optical scan
cards.  These "proofs" are easily deniable and, thus, have no value
to prove how the voter actually voted.

Likewise, electronically, there is no way that a voter could prove how he
voted, even if the confirmation screen does list all the choices that the voter
has chosen, if that screen has two buttons: "go back", "confirm", and a
suitable logic. After the voter presses "confirm" the voter sees a "thank you"
screen without any choices present. The logic canbe set up in such a way
in terms of key presses and intermediate states that even photographing
the mouse cursor on a pressed "confirm" button does not prove that the voter
did not take the mouse out and, instead, pressed the "go back" button to
change his choices.

On the other hand, photographing a paper receipt behind a glass, which
receipt is printed after your vote choices are final, is not readily deniable
because that receipt is printed only after you confirm your choices.

    To deny that receipt the voter would have to say that the machine erred,
    which, if proved otherwise, could lead to criminal charges (e.g., the
    machine would be taken off the polls and, after the polls close the
    machine would be tallied; if the electronic tally would agree with the
    paper tally, the voter would be in trouble).

Protection against providing voters a receipt, voluntary or not, is often
overlooked by those who are not familiar with election issues.  For
example, the first press release by MIT/Caltech principals after Nov/2000 said
that the solution would be to provide the voter with a receipt showing how
they voted. Later on, MIT/Caltech reformed that view and have been doing an
excellent job at what I see as a process of transforming elections from art
to science, which is a good development after Nov/2000.

Cheers,
Ed Gerck



---------------------------------------------------------------------
The Cryptography Mailing List
Unsubscribe by sending "unsubscribe cryptography" to majordomo at wasabisystems.com