The National Post

Internet Voting: E-Foolishness

by Howard Fienberg
December 29, 2000

Is Internet voting the solution to poor voter turnout and costly balloting? A group of high-tech companies think so and are lobbying both Canada and the United States to consider it. But as the business world has recently demonstrated, adding a "dot-com" to something is no guarantee of success. Those who feel Internet voting would be an improvement over the current system could be proven grossly mistaken.

So far, opposition to Internet voting has focused on issues of access, particularly the much-touted but poorly-defined "digital divide" -- the assumption that minorities and the poor are much less likely to have a computer and Internet access and would be discriminated against. What good data we have indicates that, if any digital divide does exist, it should evaporate shortly. The real problems lie in issues of security, identification and reliability.

In the United States, the Arizona Democratic party opened this can of worms by enlisting Election.com to conduct part of its March 11 primary election on the Internet. Election.com nonsensically claims this experiment proved Internet voting is safe and that its system is hacker-proof. Like most hacker challenges, this one offered little incentive that would convince a talented hacker to (a) take time off from his well-paid work, (b) risk prosecution and (c) reveal his secrets. Given the minimal stakes involved in the Arizona primary (Bill Bradley had already dropped out, leaving Al Gore the only contender), it is unlikely anyone would have bothered to hack the vote anyway.

Encryption and security measures never outpace the ingenuity of those out to crack them and the Internet is teeming with hackers of all skills and motivations from around the world. But you hardly need basic knowledge to cause trouble. Last February, perpetrators temporarily felled many prominent e-commerce Web sites such as Yahoo! and E-Bay with "distributed denial of service" attacks. Using code readily available on the Internet, they overloaded these sites with junk e-mail, routed through other people's computer systems, until the targeted sites had to shut down. Viruses, a far more dangerous breed of code, are also easy to find in cyberspace.

All software and hardware have bugs, ranging from the innocuous to the thoroughly destructive, which take a while to discover and fix. Many of us have learned to live with the dreaded "Windows has suffered a fatal error" message because we face it so often, but imagine something similar occurring during an Internet vote. Forget about vote boxes being tossed into a river, or Florida's hanging chads: Has your vote been cast or hasn't it? Should you try to vote again, or would that be illegal? Computer systems are anything but infallible, and the Internet gets tangled up on a regular basis. When either do work, they can be slow and plodding.

Our recent election showed up some of the problems in compiling an accurate voter list. Far from helping, the Internet could exacerbate these problems, because assigning and verifying secure voter IDs is difficult. Since Election.com distributes passwords using regular mail, these can be easily intercepted or misrouted. Sure, voters who failed to receive the correct IDs and passwords could probably call a customer service line and be put on hold. But who has the time? And wasn't the point to encourage voting by making it easier?

While Canada's current voting system can be disorganized and messy in pockets, enough checks and balances exist to all but eliminate large-scale fraud and mistakes. Imagine converting to a system open not only to the usual small errors, but also to error, manipulation and fraud on a massive, national scale. Moreover, we're already experienced with almost every kind of fraud in other countries' elections, so we know what to look for and can try to prevent it. But with Internet voting, we open ourselves up to whole new venues of attack. Imagine the U.S. presidential election of 2000 and worse. Of course, we would not be subject to all aspects of the Florida chaos -- it is not clear recounts are possible over the Internet. But when an Internet vote goes wrong, we are stuck with the results.

Our national ego would gain little from this kind of trailblazing. Let the Yanks forge ahead over the cliff of Internet voting, while we more carefully consider what awaits us on the other side.

Howard Fienberg is research analyst with the Statistical Assessment Service (STATS), a nonprofit nonpartisan think-tank in Washington, D.C.

[See the original article at http://www.nationalpost.com/search/story.html?f=/stories/20001229/420168.html]


return to Howard Fienberg's page