After Iowa caucus flub, can tech be trusted in elections?

'It's fine to try and improve efficiency, but you can't nickel-and-dime democracy'

An app intended to speed up reporting of election results for the Iowa caucuses has failed spectacularly, not only confusing the electorate but perhaps poisoning their feelings toward making any technological “improvements” to the voting process whatsoever.

TechCrunch staff reporters Brian Heater, Jonathan Shieber, Zack Whittaker, Devin Coldewey and Ingrid Lunden discussed the issue informally.

Brian Heater: We all agree that this is a good sign of a healthy democracy, right?

Jonathan Shieber: Totally agree with Brian here.

Brian Heater: I’m legitimately finding it difficult to discuss these sorts of things without delving into the conspiratorial. That said, I think it’s far more likely that this was just a massive fuck-up on the part of the Iowa Dems. Chalking it up to a conspiracy is honestly giving them entirely too much credit.

Devin Coldewey: But what’s the nature of the fuck-up? Fundamentally?

Brian Heater: An app that wasn’t tested at the scale of a statewide election. The more we move away from more traditional means of accounting, the more of these we’re going to see.

Devin Coldewey: There’s a chicken-and-egg problem, though. If we want to do any kind of online voting system, we’re going to need to test it. How do we go about doing that? I’ll certainly grant, however, that a first trial run on one of the most important political contests in the presidential election isn’t ideal.

Ingrid Lunden: Well, you could say that this was a kind of test. It ran alongside a manual process and it was specifically delayed and halted because the early counts were not tallying up. I don’t dispute for a second that this was a disaster and makes the Democrats — I don’t care if it’s Iowa Dems or someone at another level who called this shot — look bad. But just as I hope this won’t mean curtains for the Dems altogether, I also don’t think it’s curtains for this kind of approach of using an app to help tally and parse the results at a faster pace.

Brian Heater: I’m not a voting app expert, but there are trials for running this at a much larger scale. It boggles the mind that this was barely publicized ahead of time. Certainly that doesn’t bode well for their confidence in the project. There’s zero chance of restoring the electorate’s faith in the results after a bungle like this. The best and only thing to do is ensure this won’t happen again. Which it almost certainly will.

Jonathan Shieber: Other companies were apparently approached about building this thing and declined because the time frame was too tight and the projected contract too small to do a good job. It’s fine to try and improve efficiency, but you can’t nickel-and-dime democracy. If you want properly functioning tools you have to pay for them, and that’s something that should actually be handled by the federal government in a functioning society. There should be real funding provided by the FEC for this.

Devin Coldewey: States have authority over voting methods, though. That’s a problem for a centralized system.

Jonathan Shieber: Paper ballots for everyone. Accountable and transparent.

Devin Coldewey: It’s hard to say whether this was primarily a technical failure, a regulatory failure, a private business failure, a local government failure or what. It seems to be a bit of all of them. So it’s hard to know where to start fixing it.

Brian Heater: For starters, be upfront about the tools you’re using well ahead of the event. Let the press and public dig around. Something this important needs to be vetted, both in terms of technological limitations and the people who are involved. There’s already (I think understandably) concern that the DNC is lining up against non-establishment candidates. This is only going to add fuel to that fire.

Devin Coldewey: The lack of transparency on the origin and execution of this is pretty ridiculous.

Brian Heater: It boggles the mind that after what happened in 2016, all of these bodies aren’t more transparent. But here we are again.

Zack Whittaker: The experts I’ve spoken to — app testers and security folk — say this app wasn’t even necessary. It was to speed things up, and of course it was going to crap out at the worst possible time because there’s no evidence to show the app was tested to the degree it should’ve been. The secrecy around this app was unnecessary and led to concerns and confusion ahead of the caucus. One expert told me that nothing more than a shared Google spreadsheet was needed for this — and yet the decision to push ahead with this app defies belief.

Ingrid Lunden: If you look at what Shadow builds, it’s a lot of analytics underneath the ‘engagement’ layer (which in this case was a voting app). Given how badly the Dems played the data game in 2016 — where they seemed to completely misread what was going on, where they should campaign, etc. while the Republicans not only captured that, but did so in large part through data — I wouldn’t be surprised if data was the carrot that was dangled in front of whoever made the choice to try such an app. Do the process faster, come away with “actionable insights” to help with campaigning for the rest of the race. Ugh.

Devin Coldewey: A lot of bureaucracy is unnecessary, though. I’m not opposed to a mobile-first approach to tabulating votes… if it works better than existing methods. Maybe people don’t like Google Docs. The problem to me was the process of creation. It seems to have been both poorly and clandestinely done. Shouldn’t transparency be the first requirement and quality follows naturally?

Brian Heater: I don’t see a fully digital platform being viable in the near term. With all that happened in 2016 — between talk of Russian hacks and DNC collusion — there’s going to be a fundamental mistrust among the electorate. I think it will continue to be important to maintain a paper trail even once these fully addressable and avoidable kinks are worked out.

Devin Coldewey: That’s fair, but the fact is we already have weird, non-secure, privately developed digital voting systems that barely function being used in elections across the country (though they do sometimes offer paper trails). It seems more doable to obsolete these technologies than to abandon the entire enterprise.

An election trainer demonstrates the difficulty of casting a write-in vote on the E-Slate electronic voting machines used in many Texas counties for the November election. (Photo by Robert Daemmrich Photography Inc/Corbis via Getty Images)

Brian Heater: As someone who votes at a public school in Queens, I’m wholly familiar with the problems of antiquated hardware. Updates need to be made and the technology certainly exists, but we’re not going to get there by doing all of this behind the scenes.

Devin Coldewey: So, “the only safe election is a low-tech election,” as the NYT puts it? Because that system has proven easy to disrupt, too, with simple methods like voter suppression and gerrymandering.

Zack Whittaker: Yes, absolutely.

Brian Heater: At a bare minimum, receipts are required.

Zack Whittaker: Paper all the way. Look at Europe, it’s slow and steady. The results come out when they come out. Instead, the U.S. prioritizes speed over accuracy. There absolutely has to be a paper trail. Electronic voting machines without a paper audit trail are not safe — not even close.

Ingrid Lunden: This is pretty circular logic you have here. Technology doesn’t work because this didn’t work, doesn’t work (sorry) as an argument. For starters, there is already a ton of technology in even paper-based elections. Consider, for starters, people voting absentee by email in some places. This was a clear example of amateur hour, but I’m not sure this will be the last time you see something like this tried out and it will eventually work.

Devin Coldewey: Natasha [Lomas] also pointed out to me that in Spain there is a public-private voting tech partnership that’s both fast and effective. In that case, where can technology benefit the current systems? Better voting machines running verifiable, open-source code perhaps?

Zack Whittaker: Absolutely. there has to be verifiable root-of-trust in the hardware, open-source designs and specifications — including open-source software — with a verifiable audit trail from a paper printer. the beauty of open source is that bugs can be found and fixed easier and all it does is improve trust in what is already a process suffering in a lack of trust. The technology exists today. It’s mostly just a case of assembling the right bits of tech in the right place and making sure shit stays up to date. But honestly, I don’t think “more tech” is the answer here. Less tech works. It has worked for hundreds of years. And paper is the most reliable backup we use when things go wrong — like this time.