A Technology Too Far

Touch-screen, computerized ballots -- officially known as Direct Recording Electronic (DRE) voting systems -- are not a way station to a glorious, all-Internet future for American democracy. They're a technology cul-de-sac.

An election system should be accessible, simple, and efficient. But it must also be as secure, risk-free, and confidence-inducing as possible. Reacting to the Florida 2000 election debacle, Congress, through the Help America Vote Act, gave states nearly $3 billion to update their election systems. Great idea. But paper ballots generally -- chads or no -- were deemed suspect. The future of elections, we were told, lay in sophisticated software and user-friendly touchscreens.

For purveyors of expensive, new-fangled election machinery, it was a marketing godsend. For many election officials, federally funded DREs offered an intoxicating vision of elections too precise to be controversial.

Today, this once-bright vision of paperless voting is unraveling. In 2002 and 2004, vexing problems in state after state ran the gamut from lost and miscounted ballots to malfunctioning machines. Diebold, a major DRE manufacturer, inadvertently published its programmers' source code, betraying not only a serious security breech, but laughably easy-to-hack code.

The 2006 cycle is bringing more bad news for DRE apologists. New Mexico recently enacted legislation requiring all-paper elections -- forcing many counties to mothball DREs purchased but not fully paid for. Governor Bill Richardson noted, “Some believe that computer touch-screen machines are the future of electoral systems, but the technology simply fails to pass the test of reliability. As anyone who uses one can attest, computers break down, get viruses, lose information, and corrupt data.”

Meanwhile, the Maryland Assembly voted 137-0 to jettison the state's $90 million investment in DREs. Instead, lawmakers would use optically scanned paper ballots -- the same as Oregon's vote-by-mail system.

Never have so many citizens voiced so many doubts about the integrity of their election system. While an estimated 40 percent of all voters will cast ballots electronically in 2006, 25 states now require these DREs to produce a verifiable paper record of each vote.

But the scramble to spend billions more, upgrading, maintaining, and retrofitting DREs to generate a paper record, is a fool's errand. Such paper records are still just a receipt, whose accuracy ultimately relies on electronic data that, by its very nature, is intangible and not subject to independent audits. Better DREs are certainly possible. But in this particular war, paper will -- and should -- ultimately win.

This judgment may sound premature, even anti-technology. Indeed, since 2000 I have been an executive of Hepieric, Inc., an Oregon company that sells a wide range of software and technology-related services. My experience has given me enormous respect for well-designed software. We increasingly entrust our money and even our lives to such technology. Ubiquitous, constantly repeated commercial electronic transactions build confidence. Voting, by stark contrast, is rare and episodic. Yet the outcome of electoral transactions can be so far-reaching and profound that any notion of acceptable risk is, well, unacceptable.

It's not unlike nuclear power, and the shimmering mirage of electricity “too cheap to meter.” Human error can't be engineered away -- nor can nuclear waste. At day's end, nothing is as tangible as actual ballots, marked by actual voters, which can be touched, scrutinized, and even occasionally argued over for ultimate meaning.

And if we lack full confidence in the integrity of elections involving hundreds of thousands of machines deployed in supervised locations -- how do we get to Internet voting, with hundreds of millions of voters, entry points, and authentication issues?

Each year, corporations spend billions on hardening databases, hiring security consultants, and building in multiple safeguards to foil ever more sophisticated hackers and thieves stalking credit-card numbers and medical records. Is this the future we want, for local and state officials already cash-strapped to provide basic public services? Wouldn't the relative security of different jurisdictions' Internet voting systems vary wildly, based on technical competence and discretionary tax dollars?

Electronic voting is a technology too far. Better to heed the 18th-century wisdom of Benjamin Franklin, who so understood both the power of the visionary -- the invention of modern, representative democracy -- and the promise of the eminently practical, which included the then-revolutionary notion of creating a national postal system.

Use the post office to efficiently deliver ballots. Let voters actually mark the ballots -- and return them via the same post office (or in person). And yes, use computerized machines to count (as opposed to record) the votes. It does save time and money. But always allow the hands, eyes, and brains of real people to make any final decision. It's not a perfect system -- just the best one we can imagine, right here and now. tap

Phil Keisling served as Oregon's secretary of state from 1991 to 1999.