Some in the election administration space haven’t found a lot of other options when it comes to accountable, transparent voting systems — but new options are finally beginning to pop up.
Electronic voting machines revolutionized the democratic process — until we started to notice all the vulnerabilities, that is, and the fact that they age like any other technology. What at first looked like a farewell to millions of paper ballots turned into a nail-biting nightmare as reports of operational flaws and machines at the end of their effective lives took headlines.
After the scandal surrounding hanging chads in the 2000 election, the federal government launched the Help America Vote Act (HAVA) of 2002, which was aimed at funding the purchase of new voting technologies across the country.
Ahead of the 2008 presidential elections, Michigan’s Oakland County reported significant errors with its elections and reporting software, claiming that the M-100 voting machines, made by Election Systems and Software (ES&S), tabulated varying counts of the same ballots. Officials said that dust and debris on the sensors was the cause of the discrepancies.
Even a presidential commission tasked with reviewing the voting technology dilemma raised concerns at the highest levels with what they found in 2014: large swaths of equipment purchased through the HAVA appropriation. And by 2014 technology standards, the voting machines were “reaching the end of their operational life.”
But past all of the pre-iPhone tech waiting to experience a fatal error, some in the election administration space haven’t been able to find a lot of other options when it comes to accountable, transparent systems. Finally, however, new options are beginning to pop up. For some, the answer is open source. And as Travis County, Texas, Clerk Dana DeBeauvoir would tell you, sometimes you just have to make your own system.
For DeBeauvoir, aging voting technologies have been a point of concern and outside criticism, but she said replacing them hasn’t been as easy as one might think. While there are options on the market, none of them offered the features outlined by a county-community working group.
“What we determined was, yes, we need electronic voting, but we also needed [a system] that was evidence based, that had some sort of auditable function,” she said. “So, that was the goal, the task that they left me with: to go forth and find a system that was evidence based, and none existed. There weren’t any.”
DeBeauvoir leveraged the voices lobbing critiques at the larger system and was eventually able to gather experts to design a better election system.
“I put the challenge back on them. If you’re not happy with electronic voting as it is, then design me a system that is OK because we need electronic voting,” she joked.
The concept would be called STAR-Vote, and it would rely on a combination of electronic voting features and a printed ballot, and would be backed up by auditable reports and protected by additive homomorphic encryption.
Voters would be given a receipt with a hashcode, and unlike the current system, it would not rely on a massive amount of preprinted ballots.
“This is a different kind of approach, STAR-Vote is. For the first time, it lets us have all of the answers to prove up any election contest. We do not have that right now,” DeBeauvoir said. “If somebody wanted to say, ‘Well, how do you know?’ The answer would be: Well, I know because of all of these steps we took and all of the security measures. But the ultimate criticism of electronic voting is that it has no copy and there is nothing that can be said to that criticism.”
The county official hopes to publish a request for proposal to build the system in the next few months, and aims to have the complete system operational by 2019-2020. The next several years will be dedicated to building, testing, certifying and implementing STAR-Vote.
Joe Kiniry is a longtime election researcher and self-described hacktivist working on high-assurance secure systems with Galois, where he has been working on the open source election, public benefit corporation called Free & Fair. He also was involved in discussions around Travis County's system.
The freely available technology launched on May 16, and is designed to be useable on off-the-shelf hardware while maintaining security.
"All the work that we do on secure systems is all open source, down to the brass tacks, because the only way to have certainty about elections infrastructure is for one, to make it transparent so that anyone can evaluate it and give feedback to help make it better,” he said. “But for two, you also need to use that combination of open source and the advanced rigorous engineering technology so that you can build evidence about the system’s correctness and security that is evaluatable by anyone and by any third party.”
In his work stateside and advising European governments, Kiniry said one trend repeats in the election space with regard to computers: They are launched, and questions and concerns surface to the detriment of the democracy.
“We’ve seen all over the world the adoption of computer technology for elections," he said, "and as soon as people start peeling back the electoral system as a whole and looking at how those computers are used, who built them and what trust can we have in them, oftentimes you start to see democracies retract and stop using computers in various pieces and places."
He said he saw this pattern repeat in Holland and Ireland as they struggled with the same technologies in 2000. In both cases, the government stepped back to more transparent voting systems.
And systems without any accountability or paper trail, Kiniry said, open the door to the potential for bad actors to change the course of elections without any real proof. And it isn’t the small-time hackers that worry him, it’s the big ones like North Korea, China and wealthy interests that pose the most threat.
“At its core, the concern is that every time we look at these existing proprietary elections systems, we find that they aren’t built to a quality level that is necessary for and election," he added. "We find that they are riddled with flaws, bugs, they can’t count right, they can’t calibrate right, they have enormous security gaps and are easily subverted by experts with security knowledge. And that’s a problem.”
Despite his work to shore up the technological levies of the electoral systems, Kiniry said he is not one to dismiss the value of a paper ballot. Perhaps it’s knowing what he knows about the state of things that gives him this perspective.
“I’m one of those computer scientists that actually advocates for crazy, crazy, using paper in elections," he said. "I’m one of the 99.9 percent of security professionals that says we need to continue to use paper in elections and we shouldn't use things like Internet voting because we don’t know how to solve the problems."
Stephanie Singer, a Philadelphia city commissioner from 2012 to 2016, said in a press release that it's about time someone applied the same level of security to elections as are applied to America's defense and intelligence services.
"And because the software is open source and appropriate for commercial, off-the-shelf hardware," she added, "election administrators can save millions, if not billions, of taxpayer dollars over the next decade in implementation, maintenance and upgrade costs.”