I am very happy with my new iPhone SE, but there’s every reason not to be. Recently the Atlantic published the ominously titled “Have Smartphones Destroyed a generation?” The short answer: it sure looks like it.
Even as a childless Gen X’er it doesn’t take me much imagination to appreciate that for today’s teenagers a life without their smartphones is as unthinkable as one without running water. But aren’t we losing a sense of proportion? Running water and sewage systems have kept our cities free from cholera while Apple and Samsung give rise to new and unintended epidemics: all this ceaseless staring at your screens leads to upper back trouble, sleep deprivation, depression and a reduced interest in social and sexual relations. Conservative forces won’t see much wrong with the last one, I can imagine.
Having allowed the smartphone to permeate every pore of our existence causes Big Data no end of joy. The phone sucks up data useful to third parties in ways we can’t possibly imagine. Quantity and correlations are the keywords. Every Saturday morning around seven I instruct Siri to set an egg timer for nine minutes, so Apple knows I’m not a vegan and don’t sleep in, meaning I probably get enough sleep during the week. But this is kid’s stuff. Two years ago a Silicon Valley startup (where else?) was working on an algorithm to help establish creditworthiness by correlating just about everything your phone has registered, including your charging behaviour and whether you include surnames in your address book. They probably perfected it by now.
Apparently the odds of paying back your loan are a few percentage points better if you don’t let your battery level drop below 25% and carry a charger at all times. All this is purely based on correlation. The causal connection is flimsy at best, but then again causality is philosophically dubious to begin with. Take twenty more of such metrics and the margin of error in the AI’s decision will be perfectly acceptable to the bank. There will always be false positives (bad customers accepted) and false negatives (good customers turned down), even if it’s a human being making the decision. She will after all base her judgement not just on the rules, but also on intuition and unwitting prejudice. The algorithm may actually do a fairer job.
You’d think algorithms don’t discriminate, but they’re as (un)biased as their creator designed them to be. And the real culprits manage to stay under the radar quite well, as the ‘defeat device’ built into Volkswagen diesel engines painfully bears out. In case you’re reading this a few years hence: the fraudulent software could recognize when the engine was run in a test setup and switch the engine to reduced power consumption and emission levels.
We should never accept digital obscurity, but least of all where our democracy is concerned. This is why voting computers after a brief appearance have been outlawed in Dutch parliamentary elections since 2006. It’s not just that the system was vulnerable to hacking and snooping (it was), but that the machines were provided by a single company, with the hood welded shut. Just like my iPhone.