What all that personal data says about you
Despite algorithms, tech giants do seem to have a difficulty defining user interests
take some perseverance. Among the more than 3 gigabytes of data Google sent me (yes, that little: I kept Gmail and Google Photos out of my request and I’d recently wiped the search and browsing histories from my Google account), there was a record of my movements — a location history.
The timestamp is in UNIX time, and it can easily be converted into a normal date. Google knows where I’ve been since Sunday, February 24, 2013. One needs to divide the coordinates by 10 million to get to the accepted format, which can be entered onto a map.
The ones in my Google file point to the village of Malaya Vishera in Russia’s Novgorod region. I have never been there.
Ad topics
Unlike Google, Facebook provides a convenient interface to view the data files. With enough patience, one can discover all the Messenger conversation histories that you haven’t specifically erased — some with people you have blocked or forgotten; there’ll also be every ad you’ve recently clicked on and all the likes and comments you’re ashamed of.
But, perhaps most intriguingly — because it’s the data closest to the heart of Facebook’s business model, based on micro-targeting commercial messages to you — you’ll see ad topics with which the system identifies you. Mine are as follows:
■ Books
■ Nana (manga)
■ Forbes
■ The New York Times
■ Republic
■ Zen
■ Nassim Nicholas Taleb
■ Amnesty International
■ Angela Merkel
■ Vedomosti
■ Buddhism
■ Funny Images
I’m bothered by the idea that Big Tech thinks of me as a Buddhist who sits in Malaya Vishera and looks at mangas (who on earth is Nana, anyway? I fear that googling her may reinforce Big Tech’s impression that I’m a fan).
Then again, it may go with my psychological profile as compiled by the likes of Cambridge Analytica. The profile I get from the University of Cambridge Psychometrics Center, the origin of the research that informed CA’s methods, says, based on my Facebook and Twitter feeds, that I’m 27 years old.
I’m actually 46, so I won’t argue childishly with the psychological traits ascribed to me by the profile. Perhaps some of them fit, and perhaps they fit the manga-loving Taleb fan from the Novgorod region — surely they fit someone.
I do give up a lot of data to the digital giants. Some of it is fake, some real, some is distorted by the devices used to harvest the data (I suspect the Malaya Vishera bit falls into the latter category) and some are ruined by joint use. My younger daughter, for example, buys ebooks on Amazon more frequently than I do, because her books are shorter, and that skews the ad targeting considerably: Amazon seems to think I’m eight.
Location data
What I really want to know, however, is how exactly the tech firms use what I give them. Google says it uses the location data, among other things, to give me traffic information about my regular commute — and it does so faithfully every day as if I drove to work and back.
I don’t, I use public transportation. But I also know that if my wife were the jealous kind, she might someday demand to see specific Google location data for a certain day — and I’d struggle to explain what she’d find.
It’s not enough for the tech firms to give users access to the collected raw data, which most users will struggle to interpret, anyway. (These data dumps are so inconveniently structured that one gets the feeling they are designed to obfuscate). What I’d like is, for every type of data, a full disclosure of the purposes for which it is processed and the conclusions drawn from it by algorithms.
Essentially, I’d like to be able to push a button next to a “targeted” ad — or to any algorithmic suggestion of an interaction — that would immediately reveal exactly what information I provided to make the algorithm decide I must see it. Then, I’d like to be able to erase that information immediately.
I’d also like, with my data dump, a detailed explanation of what the company does with each data type that it harvests and stores. I just can’t think of a reason why Google would need five years of my location data, for example.
Controlling my information ultimately means just that: I want to be the editor of my data as provided to business entities, and I want much better tools for that job than the ones provided today. Google, Amazon, Facebook and Twitter know how to design a friendly user interface; they just haven’t done it here.
This may have been wishful thinking a year ago. But thanks to the CA scandal, the tide is turning and regulators are getting fidgety.
Facebook announced that they are giving users more control over their privacy through more accessible settings, which will no longer be scattered through 20 or so different pages.
That’s a baby step. If Facebook and other data harvesters don’t go much further, there’s now a real chance they will be forced to.
Essentially, I’d like to be able to push a button next to a “targeted” ad — or to any algorithmic suggestion of an interaction — that would immediately reveal exactly what information I provided to make the algorithm decide I must see it. Then, I’d like to be able to erase that information immediately.