EFF's eVoting Round Table
Tim alerted me of EFF Austin's eVoting round table so I checked it out last night. The main presenter was Dr. Dan Wallash from Rice. He was part of the team that did a scathing analysis of Diebold's eVoting system in July. The talk was pretty good, although I'd read the majority of that info on-line.
While room was filled mostly with über-geeks, there were a few government people there (county or state election types) and even a lawyer from Hart Intercivic, the company that makes the eSlate DRE (Direct Record Electronic) system that's in use in Travis County, among others.
Dr. Wallash's talk was decrying DREs in general. He pointed out some weaknesses of the eSlate and went into great detail of the Diebold system. The Diebold software was (and probably still is) a big mess, rife with security holes that would allow sophisticated "bad people" to interrupt or completely subvert an election. Nobody really knows how the eSlate works since its code, along with pretty much any other e-voting software used here in the US, is kept secret.
The list of suggestion to make eVoting safe and accurate was actually pretty short and doesn't sound too expensive. Beside the obvious, like having non-buggy code, using cryptography correctly and employing authentication protocols that actually work, the main point driven in was to have a voter verifiable paper trail. This means that when you vote, the machine would have an attached printer which prints your filled-out ballot that you can review and then drop into a ballot box. You still have the advantages of an e-system (better UI, no invalid ballots, fast tabulation) but you'd also have physical paper ballots that could be used for spot verification (count all paper ballots in select locations to make sure they match the electronic count) and for manual recounts. Another advantage is redundancy: if the computer records get destroyed, you can use the paper and vice versa.
I learned some interesting facts. Australia uses open-source eVoting software which was commission by the government--anyone is free to take a look at it and point out any potential problems. Brazil has been using DREs for a while and recently added voter verifiable printed ballots. An interesting aspect of Brazilian voting is that, due to the hight illiteracy rate, they show you the face of the candidate you're voting for. Some states, like Nevada and California, are introducing laws that require eVoting systems to have a voter verifiable paper trail. Most election officials are still pretty clueless and/or indifferent, and usually pretty defensive about the whole issue. And just because a system is "certified" doesn't mean that there aren't any problems with it (the Diebold system reviewed by Dr. Wallash's ream was certified).
Next time you vote on an eSlate, ask yourself how you are guaranteed that when you push the vote button, you're vote is actually cast, and cast for the candidates you intended to vote for.
0 TrackBacks
Listed below are links to blogs that reference this entry: EFF's eVoting Round Table.
TrackBack URL for this entry: http://www.nonplus.net/mt4/mt-tb.cgi/737
Stepan pretty much summarized what I got out of the meeting. This is something that should concern every voter, not just geeks. We need to demand that our state and local government officials hold the companies that make these machines accountable and that they hold them to a reasonable standard. There's a story in Wired today about more Diebold shenanigans, http://www.wired.com/news/evote/0,2645,61637,00.html?tw=wn_tophead_5. Also, here's a link to the actual report http://csserver.evansville.edu/~sc87/diebold/20030724_evote_research_report.pdf.