In a 2016 testimony addressing the House Committee on Space, Science & Technology, Dan Wallach warned that the country’s voting infrastructure was vulnerable to hacking by foreign governments. Computer scientists have long spoken of the dangers of electronic voting machines, and now they’re warning against adopting online voting. But is there a fundamental difference in the way that academics and entrepreneurs approach the risks and benefits of online voting? How might the introduction of online voting change the way that we conduct American elections?
SaneBox sorts through your email and moves all of the trivial stuff into a different folder so the only messages in your inbox are the ones you actually want to see. They also have a feature called BlackHole where you can relegate a sender’s messages to obscurity. For Predicting Our Future listeners, visit sanebox.com/future and receive an extra $25 credit on top of their two-week free trial.
CEO of Smartmatic
Professor at Rice University
Professor at Stanford University
VP of Elections Operations at Everyone Counts
Chief Security Officer and SVP Research & Security at Scytl
An Exercise In Frustration: Testimony Before Congress
Dan Wallach is a Professor in the Departments of Computer Science and Electrical and Computer Engineering and a Rice Scholar at the Baker Institute for Public Policy at Rice University. On September 13, 2016, he testified before the U.S. House Committee on Science, Space, and Technology for a hearing entitled “Protecting the 2016 Elections from Cyber and Voting Machine Attacks.”
Dan Wallach: “My main message for you here today is that our election systems face credible cyber-threats from our nation-state adversaries, and it’s prudent to adopt contingency plans before November to mitigate these threats.”
Professor Wallach is one of the nation’s foremost experts in researching electronic voting systems. During his testimony, he addressed the real concerns posed by Russian hacking of DNC emails, and also pinpointed where hacking might occur in the presidential election that, at that time, was less than two months away.
Dan Wallach: “I believe my top concern is the voter registration systems because they are generally online. If it’s online, it’s accessible from the internet, and if it’s accessible from the internet, it’s accessible from our nation-state adversaries. . . . If you can selectively or entirely delete people who you’d rather not vote, the current provisional voting system can’t really scale to support a large number of voters who are filling out affidavits and following that process.”
Dan suggested we invest more in computer backups and be prepared to restore data from those backups if the original voter registration lists became compromised or corrupted. He also identified a second area of vulnerability around whether our adversaries can get malware into our electronic voting machines. His solution? We must replace our aging electronic voting machines with either “next-generation” optical scan systems or new touchscreen systems, both of which would have paper trails for subsequent auditing if there were allegations of fraud or tampering so that it’s much more difficult for outside actors to manipulate them.
Dan is not a fan of the electronic voting machines we use in our elections, but at the time of the hearing where replacing the machines was not an option, his recommendation to Congress was for aggressive contingency planning in case it became evident that the machines were tampered with or disabled.
The Biggest Risks To American Democracy
While a day doesn’t go by without a news story about Russian influence on the election, Dan Wallach’s worst fears about a direct hacking of the election seem to have gone unrealized. Experts generally agree that the votes cast weren’t tampered with and that Donald Trump, while losing the popular vote, did in fact win the Electoral College, and was rightfully inaugurated in January as the 45th President of the United States.
At one point in the hearing before Congress in September, 2016, Dan addressed the issue that was most of interest to me in regards to the future of online voting.
Dan Wallach: “As a quick note, our immediate future should not include internet voting.”
And while Dan acknowledged that someday Internet voting would definitely happen, he emphatically tried to shut down discussion of it for now.
Dan Wallach: “It’s hard enough to protect the online systems that we already have. Moving additional voters online increases the risks.”
Dan described to me the enormous risks the country faced from hacking, and based his conclusions on Internet voting upon those risks. But what about the risk of a disillusioned electorate — almost half of which would stay home on Election Day less than two months after Dan delivered his testimony?
Mitigating Risks Versus Designing A Functioning System
I was really interested to talk with Dan Wallach after reading his testimony. There were two areas I wanted to explore with him. The first had to do with voter turnout, and the second related to how we might technically solve some of his concerns about the vulnerabilities of online voting. At the outset, it’s worth pointing out that Dan couldn’t point to any instance of election fraud in modern times in the United States. Of the two examples he did cite of election fraud (one in Mexico and one in the Ukraine), the former fraud occurred with paper ballots.
Typically, when you hear a technologist speak about the requirements for building any system, they begin by articulating all of the objectives that the system is being designed to affect. If you build a social network, you don’t just describe the requirements as allowing users to functionally create a profile and interact with others. You would actually describe your objectives for the number of subscribers you hoped your system would register and attach quantitative metrics to the types and frequencies of their interactions.
For Dan’s testimony, I imagine he was only asked about mitigating threats to voting systems, so nowhere in his testimony were the objectives of any system that might be designed to improve the voting process. I asked him about what he saw as the purpose of these voting systems: namely, whether it was simply to make sure that every vote was counted accurately, or that we also designed a system where the maximum number of people voted. He acknowledged the value of an interested electorate and was adamant that online voting was not the only way to achieve it.
Dan Wallach: “In the Dutch election just a couple days ago, they had over 80 percent turnout and they voted on paper. Voter turnout is a great thing to have and achieving high voter turnout has relatively little to do with the technology that you’re using. Achieving high voter turnout can be done in many different ways. The way that most overseas and military voters vote today actually is that that you can download a blank ballot, print it, and then mail it back. So that way, you have the postal mail as only for one direction, rather than both directions. . . . So there are other solutions besides full online voting.”
Dan made clear he recognized the benefit of having all interested voters participate in an election, although he suggested that making turnout of all voters a primary objective for a voting system might not be all that desirable.
Dan Wallach: “A political scientist who’s a colleague of mine likes to make the following quip, which is: in order to vote, every voter has to have something that they all have in common, and that something is that they’ve decided who they want to vote for. And if you push to increase voter turnout, you’re increasingly getting people who have less strong opinions about who they’re voting for. At the limit, you’re getting people walking into the polling place with no idea who they’re voting for. So be careful what you ask for when you’re trying to increase voter turnout.”
Alabama: Evidence For Online Voting Leading To Increased Turnout
As I conducted research for this podcast series, I found some evidence in the United States that Internet voting can dramatically improve voter turnout. James Simmons is the VP of Elections at Everyone Counts, one of the worldwide leading hardware and software providers for voting. Everyone Counts conducted the first Internet election in the United States at scale. They worked with the state of Alabama to allow Alabama servicepeople stationed overseas to vote directly over the Internet for the Presidential election in 2016.
James Simmons: “Alabama worked with us to allow their military and overseas voters to vote online using our platform. So a very small segment of the U.S. population did get to vote over the Internet. And so we can look at that segment because we understand the statistics behind that.”
Me: “Was that the first time that any segment has voted over the internet?”
James Simmons: “Yes, in the sense that it’s the first time someone has logged into a web-based application, filled out their ballot, and clicked submit. There have been a number of cases of folks basically sending and receiving their ballot by e-mail, or filling it out and then uploading it into an SFTP site type thing. But in terms of truly engaging with a web application, yes, this is the first time in U.S. public elections that that’s happened.”
Me: “What was their percentage of participation in prior elections and what was their percentage of participation in this election?”
James Simmons: “For the 2012 presidential election, the participation for that demographic was just under 13 percent. So it’s hard for these folks to vote. Getting a mail ballot out to them and getting it back in time through international post, especially in some of the places that these folks are serving, was a challenge. For the demographic that we supported there, the 2016 [group] saw an 84 percent turnout. So that group of people — military overseas folks — beat our national average by a pretty significant margin.”
Me: “So why Alabama? First of all, why doesn’t this exist nationwide? Why in this past election cycle did we not have that same opportunity?”
James Simmons: “A lot of that is just the willingness to be the first to try. And that’s I think the answer to your first question of, ‘Why Alabama?’ They recognized that they had this group of folks who were underserved in their ability to participate and they chose to take a leap with us — first to pilot it during their primaries, and then to go for it during the general election — and they were the first to do so.”
Approximately 4,000 Alabama servicepeople participated in the 2016 election. I was wondering if there was a single example of where Internet voting could improve turnout by a few percentage points. Here, participation increased by over 70 percentage points.