What computers know about us
As software monitors our online behaviors, the possibilities for misinterpreting what’s going on, not to mention invading privacy, seem legion.
The electric company sends me colorful reports on my monthly energy habits. A recent one rated me “above average” in controlling electricity use as compared with my supposedly more extravagant neighbors. I’m usually “average.”
I would have basked in such praise were it not for this pertinent fact: I wasn’t home that month.
Thus, I had not flicked a single light switch, touched the microwave or even recharged a cellphone in the period being monitored. The only things driving up my kilowatt-hours were the automatic-drive appliances — the refrigerator, the cable box, the digital clocks. How come I didn’t qualify as “well above average”? I have no idea.
This is an especially primitive example of computers crunching personal data into faulty conclusions. And seeing as judgments are being made, shouldn’t the utility factor in the number of people residing in each house? If a family of six lived next door and used twice as much electricity as I do, I’d consider the neighbors thriftier than I.
A recent phenomenon called digital phenotyping makes even fancier (and more fanciful) claims for the benefits of computers tracking human behavior. Promoters say they can assess one’s physical and mental health by scanning his or her use of social media — findings that can be shared with doctors.
The possibilities for misinterpreting what’s going on, not to mention invading privacy, seem legion. But the theory is interesting.
For example, someone posting a slew of tweets at midnight may be suffering anxiety. It has been noted that late-night tweeting spikes in times of political turmoil. Posting at 3 a.m. might suggest a sleep disorder. Might. Tweeting madly in the wee hours could also mean someone has customers in Asia.
Few things should provoke more anxiety than the idea of Facebook’s using its data to dig into our health status. For example, the company is now employing artificial intelligence to scour posts in search of suicidal thoughts. Friends responding with comments like “Can I help?” add to the suspicions that someone is in serious trouble. Facebook review teams have been known to alert local authorities on the basis of troubling patterns.
There’s huge irony in Facebook’s presuming to help people racked by anxiety when the social media giant has become a major source of it. Facebook, after all, gave an enormous platform to shady foreigners trying to stoke anger and resentment across the land. Do you want a Russian troll impersonating an American friend at a time of stress? (For more insight, watch “The Russians.”)
Many psychologists already blame social media for rising rates of depression. Facebook and others have replaced intimate physical friendships deemed key to good mental health. Also, the presence of Facebook friends showcasing their nearly perfect lives (whether real or not) has been found to foment feelings of envy and inadequacy in others.
In other news, a “digital health” company called Sharecare purports to measure stress levels during phone calls. Pattern recognition technology lets its computers send out messages such as “you seemed anxious.”
Having grown up in an environment where anxiety was a base-line reaction — “The bird feeder’s empty!!!” — I worry how often some people are going to be beeped, buzzed or messaged. Also, aren’t beeps, buzzes and instant messages themselves a source of stress?
My distrust of this technology notwithstanding, I admit to voluntarily letting my Fitbit watch harass me whenever I sit too long. But here’s the truly creepy part: When I get up, take the 250 steps and see the watch face light up with “Nailed it!” I actually feel lauded.
Where this is leading, no one knows. That we can still unplug from it remains a source of comfort.