A Not-So-Short History of Electronic Voting
As we close in on the 2020 elections, here’s an in-depth look at the history of electronic voting. Originally published on Spiceworks in 2016.
It was 3 a.m. and Jane Platten was staring into a room full of electronic voting machines in Cleveland, OH at the end of a 22-hour day. She pointed mutely over at the machines, which more than 200,000 people had cast their votes into that day.
Earlier that day, the machines had worked as promised, recording electronic records of each person’s vote on a memory card. These records would be transferred to what Platten and the other workers there called the “GEMS server,” which in turn produced an official tally of votes.
The stress started when that server stopped working. “Cuyahoga County technicians clustered around the computer, debating what to do,” The New York Times wrote. “A young, business-suited employee from Diebold — the company that makes the voting machines used in Cuyahoga — peered into the screen and pecked at the keyboard.”
Several tense minutes passed, but no one could figure out why the machine wasn’t working. The technicians eventually turned the server off and on again. Once more, the machine started tabulating vote counts and work resumed. But an hour later, the server froze up again. Once more, the technicians were left turning the machine on and off again.
“When the votes were finally tallied the next day, 10 races were so close that they needed to be recounted,” The New York Times says. “But when Platten went to retrieve paper copies of each vote — generated by the Diebold machines as they worked — she discovered that so many printers had jammed that 20 percent of the machines involved in the recounted races lacked paper copies of some of the votes.”
Platten’s story is not isolated. Reports exist about electronic voting machines “flipping” votes, getting vote counts wrong, and suffering from poor security protections. Designed to make voting easier and counting ballots faster, early proponents of electronic voting machines touted how these machines further democratized elections, reducing the difficulty of getting paper ballots to every corner of the country. In short: The origins of electronic voting machines were pure.
So what are those origins? What might the future of electronic voting look like? And should we be concerned about a DDoS attack like the one against Dyn in October 2016 disrupting future elections?
The rise of e-voting machines
In the 2000 presidential election, America’s voting process made international news after problems with voting machines caused a last minute recount in Florida, which ultimately led to a Supreme Court decision to break the deadlock between George W. Bush and Al Gore.
To prevent future voting problems and stave off potential embarrassments, the US government passed the Help America Vote Act (HAVA) in 2002. Among other things, HAVA sought to replace punchcard and lever-based voting machines by giving states $3.9 billion to invest in more modern equipment.
“The federal largesse led states to begin buying Direct Recording Electronic (DRE) voting machines from companies such as Sequoia and Diebold,” The Verge says. “Early DREs left no auditable paper trail: votes were recorded directly to digital memory.” And as time went on, reports emerged about problems with machines. With no paper trail to verify that the votes had been correctly counted, some election officials were left unable to confirm or deny such reports.
Soon, worrying reports began to emerge about security flaws in voting machines such as WINVote, which one computer scientist said had only been safe because no one tried to hack them. Plagued by serious security flaws and an insecure Wi-Fi feature to tally votes, WINVote machines were employed in Virginia alone for over 12 years (the state decommissioned the machines last year).
But here’s the thing: Despite warnings from computer scientists and software developers, there is almost no evidence any voting systems have ever been successfully hacked. In fact, experts will frequently say the real danger isn’t hacking, but error.
“If there are guys who are trying to tamper with elections through manipulation of software, we would have seen evidence of it,” says Ed Felten, a computer scientist at Princeton. “Nobody ever commits the perfect crime the first time. We would have seen a succession of failed attempts leading up to possibly a successful attempt. We’ve never seen it.”
This is where things get strange. While state officials cite hacking as a reason to replace faulty touch-screen voting machines and media publications stoke fears that a hacker could turn an election, the real concern is more basic: Will the machines themselves work?
There’s no such thing as 100% uptime
The truth is no computer system has 100% uptime. And the irony surrounding machines that were supposed to remove the paper ballots and the arduous counting process to make things easier? It turns out electronic voting machines that produce a paper trail are actually easier to deal with.
After all, proving that an election has been fairly settled is a lot easier when you can show the losing side a paper trail. Most states do perform post-election audits to verify vote counts. But a small number of states that use paperless DRE voting machines have no audit mechanisms in place and no paper trail to audit in the first place.
“The problem with putting these auditing systems in place is the same one keeping more reliable voting machines from the booths in the first place: a lack of money and political will,” Wired says. “There’s new voting equipment out there that’s much more secure than the machines states purchased in bulk a decade or more ago, but only a handful of states and municipalities — Rhode Island, DC, and parts of Wisconsin among them — have upgraded in the past year.”
“I think the early enthusiasm for this technology has faded and what we’re seeing is a shift toward what should be the next phase, which is hybrid systems combining an electronic record and a paper record,” Alex Halderman told The Verge back in 2012. “This really should be seen as an advance and not as a reversion or anything like manufacturers were originally painting it. It’s really a big security improvement over both paper and purely electronic systems.”
The rise of optical scanners
But there’s one type of electronic voting machine that stands above touch-screen voting machines in terms of security and efficacy. It’s the optical scan voting machine, a piece of technology that was originally manufactured en masse to help grade college entrance exams. Just like with standardized tests, citizens cast their votes “by filling in an oval, box, or similar shape on a paper ballot.” A machine then reads — and tabulates — the ballot.
The irony, of course, hits you over the head: The basic technology that makes these machines work has been around for years. Progress it seems is sometimes intelligent regression.
Unlike with touch-screen voting machines, votes and ballots are “immediately tangible to the voters; they see it with their own eyes, because they personally record it.” And if a recount or audit needs to conducted? There’s a paper trail that’s easy to access.
Some experts say the use of optical scanning machines leads to “perfect elections,” where voters are confident in the results and candidates are able to confirm the accuracy of the final count. Speaking to The New York Times in 2008, the Florida election official Ion Sancho said his error rate using optical scan voting machines was “three-quarters of a percent at its highest, and has dipped as low as three-thousands of a percent.”
But Sancho raises another important advantage of optical scan voting machines: It’s trustworthy. “In one recent contest, a candidate claimed that his name had not appeared on the ballot in one precinct,” The New York Times says. “So Sancho went into the Leon County storage, broke the security seals on the records, and pulled out the ballots. The name was there; the candidate was wrong.”
The candidate ended up apologizing. As Sancho points out, you couldn’t do this with touch-screen voting machines. “I like that certainty,” he says. “The paper ends the discussion.”
What happens if the internet breaks on Election Day?
For all the reported problems with electronic voting machines, there isn’t a significant risk when it comes to US elections, which are decentralized and difficult to hack on a grand scale (plus, only four states use DRE machines without any verified paper trails). But what if a massive DDoS attack like the one against Dyn in October 2016 were to strike on, or after, Election Day?
“The good news is that even if the Internet goes down on Nov. 8, your vote will still count,” The Boston Globe says. “Few states use computer-based voting machines, and even those machines aren’t online.”
The takeaway here is that a large-scale DDoS attack wouldn’t affect the election — but they’d certainly gum things up. Just like those who were unable to access Twitter and Netflix last month, election officials might be unable to check voter registration information against a database. This wouldn’t prevent individuals from voting — hello provisional ballots — but it would make for longer lines.
Other more mundane things might happen, too. People might, for instance, have a tough time getting the polls without being able to look up directions online.
Then there’s what happens after the voting is over, which is arguably more ominous. “Usually Americans tune in for the quadrennial cable-news election night ritual, celebrated with two-tone maps and portentous pronouncements about the political future,” The Boston Globe writes. Not so without the Internet. You could still watch television, of course, but a lot of those election-night returns are collected through the Internet.” In other words, you’d have to get a good night of sleep before the results rolled in.
The bottom line: Don’t worry about a DDoS attack doing anything other than making the wait a little longer. Worry about the people who might complain about what that wait means.
The 30,000-foot view
For all the problems with electronic voting methods, counting millions of votes is long, arduous work undercut by the risk of human error. And not all electronic voting methods are created equal — some are decidedly more trustworthy than others.
While officials continue to explore electronic voting methods such as optical scanners or a new form of “crypto-voting” that’s being spearheaded by a Spanish startup, it’s useful to have a little perspective. Up “until the introduction of what was known as the ‘Australian ballot’ in the 1880s,” History Extra says, every vote was a public affair.
Developed in the Victoria colony in Australia, the so-called Australian ballot was a private, generic ballot that citizens filled out. Before that, some counties saw their citizens publicly yell their vote with a “yay” or “nay.” Some political campaigns would create their own brightly colored ballots and give them to their supporters.
“Voting against the prevailing mood in one’s own precinct took courage, often physical courage,” History Extra says. Violence was common and bribes were frequent. Legal battles over voter suppression and bribery in the 19th century took on practices such as “cooping,” where campaigns would ply a group of men with alcohol and then take them to the ballot box when they were semi-conscious.
In short, today’s electoral system is far better than it has been. But while states, counties, and the federal government continue to look for better solutions when it comes to electronic voting, today’s electoral system stands to see future improvements as well. After all, when voters are still mailing in ballots in the 21st century, there’s probably a better way.
Aaron Winston lives in Austin, TX and has written about technology, history, e-commerce and more. Currently, he works as a Content Strategist for the flexible workspace company Hana.
