The 2016 election has intensely focused on the debate surrounding the NSA's endless amount of spying powers. But when Iowa voters recently voiced their opinion on who should be in charge of that murky world of cyber surveillance (among other things), they didn't know they were already targets themselves.
A company called Dstillery used Mobile Ad IDs to figure out the consumer trends of people caucusing in Iowa. The method couldn't ascertain the habits of voters in highly contested counties with any reliability. The Iowa counties that were clean sweeps for certain candidates offered an alarming amount of detail about who voted for whom -- and how they like to spend their free time.
Kashmir Hill explained how the whole process works in a Fusion post this week:
Dstillery gets information from people's phones via ad networks. When you open an app or look at a browser page, there's a very fast auction that happens where different advertisers bid to get to show you an ad. Their bid is based on how valuable they think you are, and to decide that, your phone sends them information about you, including, in many cases, an identifying code (that they have built a profile around) and your location information, down to your latitude and longitude.
So all Dstillery had to do in Iowa was target specific voting locations and analyse the ad profiles of voters at those polls, using machine learning the company is able to figure out what kind of voters are at the polls. Dstillery recorded 16,000 voter IDs which is only a fraction of the amount of people who turned out to vote last week. However, that was enough to discover that people who "liked grilling" and DIY house work tended to support Donald Trump and that expectant parents often voted Republican. Strangely enough, people who liked NASCAR would either vote Trump or Hillary. Whatever that means!
Reports from the NPR and USA Today seemed at ease or relatively impressed with Dstillery's approach -- which is clever in a way. But the ability to massively categorise a group of people based on their location and ad profiles feels profoundly disconcerting, even if it is more common than you might think.
Dstillery is quick to add that correlation isn't causation and that all this data is totally anonymous, so we're not talking NSA-level, constitution-violating smartphone tracking here. But companies and candidates can figure out what kind of voter you are, without your knowledge or consent.