Okay, There’s An App For That. Why?
Ho boy. The Iowa Caucuses were…they happened. And they happened in a way that was both completely unprecedented and utterly, exhaustingly predictable. It’s a story line that’s becoming, frankly, a little boring: people use old process, someone decides that an app (or AI) will be better, lots of product development takes place, very little field testing or troubleshooting occurs, new technology rolls out to great fanfare, and then disaster. It happens when apps try to go operational, it happens when AI is used to analyze datasets, it happens when companies decide to shift their focus. It’s a slow moving catastrophe, an avoidable one at that.
The situation in Iowa is a perfect example. After the 2016 election, the Democratic Party concluded that it needed to up its digital game, and the 2020 primaries were to be the rollout for some of the new tools and methods. But, in an outcome that should surprise no one familiar with the operational dynamics of large organizations, nothing happened after the initial decision. It wasn’t until very late 2019 that Iowa Democrats hired a company — and I couldn’t make this up — called Shadow to build a caucus results reporting app. I guess Bad Omen LLC was busy.
In any case, the Iowa Democratic Party completely outsourced the development to Shadow, didn’t field test it, didn’t deploy it until it was actually the time for voting, didn’t give users the chance to see how it worked, didn’t provide tutorials on how to download and use the app, and didn’t work with Shadow to identify bugs. So, when Iowa precincts couldn’t figure out how to get the data from the local gym to the main tabulation centers, results stopped coming in, and no one had any idea who had won the caucuses.
Don’t Get Judgmental
The simple thing to do here would be to blame Shadow for the rollout and the miasma that was Tuesday night, or perhaps to blame the DNC for not properly following through on the rollout of an app that governed the first portion of what Democrats are calling the most important election of our times. And, in this case, we should absolutely do the simple thing. It beggars imagination to think that neither Shadow nor the DNC understood the stakes, or that they couldn’t have predicted difficulties in the first-time use of an app by thousands of people in a high-pressure situation. And, given that there are no signs of outside interference or hacking, this appears to be a classic instance of last-minute thinking and last-second rollout.
That said, it’s also possible to see a broader problem in this scenario — the risks of tech solutionism. Tech solutionism is the faulty belief that “technology is going to save us,” or that “technology is always the answer.” First of all, “technology” in that context doesn’t really mean anything, because “technology” actually means and skill, craft, method, or tool that allows us to accomplish a goal. Electricity is a technology, so is science. When we say “technology” what we really mean is Information Communication Technology, or ICT, and the apps/programs that we operate via those ICTs.
Tech solutionism is one of the more widely-held manias these days, manifesting itself in the idea that there is always a digital (or, ideally, cloud/SaaS) answer to every problem. That’s why the IoT is reaching places that make absolutely no sense, like the sensor-enabled diaper that tells you when your baby needs to be changed. That such a product exists raises a very significant question: why? If the analog alert system (in this case, a combination of auditory alarm (“Waah!”) and olfactory alert (“Ugghh!”) has worked well for thousands of years, why exactly are we interposing sensors and tracking into the mix? Oh, and there’s a camera and microphone involved? Good good, because there certainly haven’t ever been negative consequences to placing cameras and microphones in baby’s room.
The Shadow app is case in point. One of the central concerns in the 2020 elections is foreign interference in the process, and vote tabulation is one of the most sensitive aspects of election security. Shadow, and the Iowa Democratic party, have issued repeated assurances that there is no evidence of hacking or that the votes were compromised, but that’s all besides the point: the appearance of a flawed process is just as damaging as a hack would have been, or perhaps worse. At least with a hack this wouldn’t be our own fault.
The Analog Way
More broadly, though, the question is this: why were we using an app for a process that could be more reliably verified through human activity? As the CEO of Shadow himself said, the results of the caucuses are all publicly available and recorded on camera, so there’s no risk of a false result, or of tampering with the outcome. So what’s the need for a rushed, half-built system without the proper vetting?
A substantial part of the problem is the tech-solutionist mantra that “there’s an app for that.” Increasingly, it’s true: the millions of apps available for download or free purchase are a testament to the imaginative power and drive of their creators. But creativity doesn’t necessarily translate to good execution skills, which is why users delete 71% of their apps within 90 days. That doesn’t even touch on the legions of apps that serve largely to siphon personal data or install malware. It’s the same with AI: people talk about using it all the time to solve problems but rarely explain how machine learning will solve the problem better than a human, or whether there’s a problem to be solved at all. That’s how Barnes and Noble ended up in an (entirely preventable) mess about race, AI, and famous works of literature causing them (entirely merited) bad press.
I know what you’re thinking. Contra principia negantem non est disputandum! (You need to ease up on the Latin, okay?) Yes, it’s true that before you can make an argument you have to agree about the facts and the principles at stake. But I do agree that apps are useful and I don’t think we need some kind of return to the pre-digital, analog age. What I’m saying is somewhat more nuanced: sometimes, analog just happens to be better at delivering what’s important to us. In order to determine that, we have to figure out what’s important to us in the first place.
Put it to the Test
Consider elections. Before anything else, we have to decide what’s the most important thing about an election. I would argue that the essence, the central purpose of an election is the selection of the candidate the most people voted for. Are there other aspects of elections that matter? Sure: making sure they’re open, making sure they’re fair, making sure they’re held in an orderly fashion, etc. etc. But these characteristics are secondary, in that they support the primary purpose, and don’t form their own distinct purpose for the election. So, when it comes to running an election, the most important question is “how do we make sure that the person elected is the person for whom the most votes were cast.” Everything else comes after that.
To illustrate, compare proposals for e-voting against paper ballots. They both tally votes, they both facilitate elections. Apps, however, are faster. Does that make them better than paper ballots? Maybe, but how? Can the app be hacked? Can the app crash? Can it send corrupted results? Yes, yes, and yes. What’s the drawback of paper? It requires a manual count, which means that tallying votes can take longer — hardly an overwhelming drawback in the face of foreign interference in elections. (And yes, we know that there can be hanging chads, but, seriously, that was in Florida, and nothing from Florida should ever be an example in a serious argument).
“Obviously, this means that paper ballots better advance the primary function of elections, so let’s run out and smash all the electronic voting machines!” No, it doesn’t, and no, let’s not. Paper ballots have their own drawbacks (logistical, ecological, Floridian, etc). The point we’re making here is that there is no immediate answer to the question. It takes a serious consideration of the issue, a recognition of our values, and a critical evaluation of whether our values and actions match up with what we want to do.
It’s the same for a business thinking about an app, deciding whether to use AI, or choosing how it wants to manage privacy and data. You can always move fast and break things: it’s almost the easier way to do things these days. But does what you’re proposing advance the central aim, the essence of what you’re trying to do? How can you be sure that adopting the newest or most tech-solutionist approach will actually support your central purpose, and not undermine you? Everyone wants to think that their decisions are based on sound reasoning; the issue is honestly evaluating whether they are. And everyone wants to succeed; the issue is how. Should we dump all voting apps and go back to paper, or should we institute universal online voting? I can’t answer that question, certainly not on my own. But, desperately, we need to come to conclusions about these matters. Tackling these questions is hard, but it’s what responsible people do when they really engage with the things that matter to them, like their livelihood, their ethics, their democracy. Finding the answers takes time, and despite our wishes, there’s just no app for that.