There’s a common saying amongst those people who deal with information security: “The only truly secure computer is one that’s turned off, unplugged from all networks, encased in concrete six feet below the ground in an unknown location, guarded 24/7 by elite mercenaries who shoot first and ask questions later, and even then, I wouldn’t put money on it.”
Technological advances are believed to be a good thing, whether it be advancing from a scrub-board to a washing machine or from an abacus to a CRAY Super Computer. Progress often allows us to do things faster and more efficiently, so we begin to assume that all technological change is inherently good and neglect to note its consequences, especially when it doesn’t work.
One such technology is electronic and online voting. Last week, the University of Calgary Students’ Union held their annual by-election to fill positions left open after February’s general election. In recent years, the SU has moved from a paper ballot system to an entirely online one, with mixed results.
The argument is that online/electronic voting systems not only reduce costs by eliminating the need for paper ballots and people to hand-tabulate the results, but they also increase voter turnout by making the system more accessible. Yet, during the last election people were unable to vote for the entire first hour, causing chief returning officer Mike Brown to shut down the link and revert to paper ballots for the next hour and a half while the company that handles the SU’s elections, CanVote, sorted out server problems. In 2002, nobody was able to vote for the first day due to server issues, causing voting to be extended a day. Then again in 2004 there were more irregularities, prompting the SU to change software providers.
Granted, SU election problems haven’t been earth-shaking, but the ramifications of online voting run deeper. In the 2004 U.S. presidential election, many states used electronic voting systems manufactured by Diebold Election Systems. These are much less complex than the systems required for online voting because votes can only be cast at specific physical locations and not in say, Bulgaria. Yet there were still massive problems. For instance, during the presidential primary in California, two counties experienced problems with the voting-card encoders supplied by Diebold, disenfranchising massive numbers of voters and leading California Secretary of State Kevin Shelley to ban the use of electronic voting machines in four counties. He also recommended criminal prosecution of Diebold.
That’s just one example; there were other instances where voters would go to the machines only to find a candidate was preselected or that their choice disappeared after pressing the submit button.
There are two major issues here. First is the fallibility of technology. Quite simply, the more complex a system is, the greater the possibility of failure. Paper ballots have worked for so long because it’s difficult to screw up checking off a name on a piece of paper and then putting it into a sealed box. Compare that with online voting, where you have to verify the person placing the vote is the person behind the keyboard with a SIN or student ID. You then have to process the vote using some kind of web-based form, encrypted and sent to a redundant array of servers designed to concurrently replicate each other’s data in order to prevent information loss in the event of a catastrophic system failure. If this all sounds overly complex, that’s because it is.
The other problem is human fallibility, which has always been a major issue, even before eVoting. Whether you have an insanely complex electronic system or simple paper ballots, the information will be handled at some point by people. Thus it becomes a question of trust: who is reliable enough to handle such important information? The answer is: as many people as possible. It’s much harder to get away with something when you have the rest of the electoral district association looking over your shoulder. Yet electronic voting requires highly-specialized technicians to administer it from the top, leaving it open for tampering.
The scenario is especially disconcerting since the majority of these voting systems don’t leave an independent paper-trail, so people without masters’ degrees in computer science will have trouble understanding the results. Some systems like the one provided by CanVote can produce a very detailed paper recount of the results, but because this comes from CanVote itself we’re still at the mercy of a select group monopolizing the information.
It’s not that a conspiracy exists, it’s that the system could allow for one. It wouldn’t make sense for any group to rig a student election in which only 1,400 students voted, however, as we move more and more towards electronic voting systems in North American society, we have to ask ourselves fundamental questions: Are the speed, cost and accessibility provided by electronic voting systems worth relying on for-profit groups? And how legitimate is our concept of democracy if we’re willing to risk it in the name of efficiency?