TapeVoting

I haven’t been able to find any decent report of the electronic voting trials taking place today in the local council elections, the best is probably Jason Kitcat. I have these comments to make about the principle:

It is not presently (nor in the foreseeable future) possible to construct a secure, Internet-based system for remote electronic voting.

Dr. Rebecca Mercuri, Bryn Mawr College, 2002

The main reason is that you have many conflicting and contradictory requirements. You need to check that the mechanism to vote is actually available; the entitlement of the person to vote; that they vote only once; that privacy is maintained; that no coercion has taken place; that the voter gets positive feedback that their vote has been cast as they directed; and that the candidates and other observers get an unambiguous assurance that count mechanism is accurate and unbiased. Note that some of these are not the same requirements as for commercial transactions; that interaction is deliberately not anonymous (else you won’t get anything delivered nor charged), nor are the requirements for all elections the same.

No voting system is going to meet all these requirements but the added factor in remote electronic systems is the possibility of automation generating sufficient mis-votes to influence the outcome. Proving identity is not done at the ballot box but the attendants are going to notice gross abuse; Privacy is weakened by numbered ballot slips but it takes a manual, obvious and difficult cross reference to trace back each vote, unlike electronic systems where the identity and the vote cast can easily be in the same or linked databases; no one can twist your arm when marking your cross; you put the slip in the locked box personally; representatives of all interested parties can see the count, where the actual voting slips are present laid out on the table and they can oversee any queries that arise.

Introducing the internet into this is to shroud the whole process in a dense fog. You cannot rely on the security of the entry device (home PC) nor the transport mechanism (ISP to global internet). No amount of encryption can compensate for the huge number of home systems that are vulnerable and exposed. It is analogous to leaving ballot boxes unsupervised on street corners for a few days as you have no way to tell how the voting slips arrived. To continue the analogy, how can the voter recognise a genuine ballot box—read “spoofed voter web sites”. Finally, if you get your vote to the correct system, the opportunities for that server, connected to the world, to be attacked are not insignificant. In a recent case, personal details of applicants for NHS positions were exposed alongside their names; this is despite the system requirement to strip off these details before recording the data at all.

There are arguments in favour of electronic polling stations but the systems used must be independently audited (not proprietary black box systems) and provide a printed feedback confirmation of the vote cast which can be deposited in a ballot box in case a manual count is needed e.g. in case of system failure, compromise or dispute.

Dr. Mercuri goes on to say

To say that “it is probably impossible to make any system perfect” and then use this as an excuse to impose a horribly imperfect and flawed process on the voting public, is sorely misguided.

4 Responses to “eVoting”

^ Top