Emotional analytic apps to let your smartphone choose the company you keep.
We tend to invest technology with boundless possibilities as we become more aware of the constraints on our physical lives. It’s why technology can simultaneously promise instantaneous connection to hundreds of people while at the same time make a satisfying connection feel impossible. Which is not to say those experiences aren’t authentic, just that you can have hundreds of Facebook friends and still feel isolated or lonely, and our pre-internet friendship rituals are no longer fit for purpose.
For the past couple of months, I’ve been experimenting with pplkpr (pronounced “people keeper”), an app that claims to “automatically manage your relationships so you don’t have to”. The app imports a list of friends from your Facebook profile, asks you how a friend makes you feel before and after you meet them, and syncs up with a Bluetooth-enabled wearable device to monitor your heart rate at similar intervals. From this mix of biometric data and self-reporting it will divine trends and prompt you to take certain courses of action. It might be found that a friend arouses you, and the app could suggest a date. Another friend might make you anxious, and the app could suggest you avoid their company. A friend might anger you, and the app could invite you to de-friend them from your social networks.
Owing to my paranoia about allowing companies to have access to my vital statistics, I elected to manually record my heart rate – the wristband device isn’t essential to use the app, though it would admittedly enrich the experience. The app’s emotional barometer comprises six adjectives: a person might make you feel relaxed, anxious, aroused, angry, excited or bored. Applying these anodyne labels to my friendships reminded me of the multiple-choice quizzes often found in teen magazines that promised to detect “toxic friends”. While the magazine quiz was a handy diagnostic tool, helping muddled teens decipher the signal and noise of another human’s behaviour, pplkpr tries to interpret the latent signals within our own physiology and prescribes correctives for our digital and real-world lives.
On a recent Friday night, before meeting a close friend, Ben, at the pub, I measured my heart rate and indicated on the app that I was “excited” to see him. Returning home after a pleasant evening, I recorded how I felt. (Frustrated by my limited options for adjectives, I decided that “excited” would suffice in lieu of “intellectually stimulated and a bit tipsy”.) The app then asked me if Ben made me “most excited” or “least excited” when compared with my other friends who had also made me “excited” and, after I had ranked my friends in the app’s arbitrary excitement hierarchy, did
I want to let him know?
Self-quantification is one thing, but I wondered if quantifying (and therefore surveilling) a relationship without a person’s knowledge transgressed several social or moral codes. I took a screenshot and texted it to Ben. He did not respond. (We’re still friends.) Since then, our subsequent catch-ups have garnered more effusive responses; the app has therefore suggested I might like to message him or “hang out” with him more. Logging relationship data in this way felt like an intimate act – like keeping a diary – even if rationalism was its animating impulse. In some ways pplkpr felt more unvarnished because the data was not subject to subtle editorialising. Yet it also felt more laborious.
I had an urge to use the app when I met friends I hadn’t seen in long periods of time, or on people I’d just met, or work colleagues. I saw these events as opportunities to deploy different adjectives. I felt anxious before meeting one friend but relaxed afterward, and it appeared to be evident in my heart rate. I wondered whether marking a friend as boring would be an act of betrayal or an expression of radical honesty, and whether pplkpr’s algorithm could discern the difference.
Certain aspects of human behaviour have already been outsourced to algorithms. Dating sites such as Tinder, OkCupid and Match.com use mathematical formulas to maximise the pool of partners with whom we are most likely to have the greatest relationship potential. The algorithm’s predictive powers could be said to “optimise” the sexual marketplace, although many may feel differently. A Hong Kong venture capital firm, Deep Knowledge Ventures, last year appointed an algorithm called VITAL to its board, to supposedly enhance the board’s capacity for independent decision-making.
Pplkpr’s uncritical embrace of data trends and its seemingly facile taxonomy of emotional experience might make people bristle, but that’s the point. After all, it’s not just an app but a deliberately provocative art project designed by Lauren McCarthy and Kyle McDonald, two programmers and artists-in-residence at Carnegie Mellon University in the US. The app’s name appears to be a dig at Silicon Valley’s penchant for the arbitrary eradication of vowels.
“The idea of an algorithm tracking and managing your social life feels creepy, but what if it actually works?” McCarthy says. “What if it actually improves your relationships and emotional life? We’re interested in this dissonance we feel.” Pplkpr wants us to ask questions not only about the connections we form with people but of the nature of relationship we have with technology and ourselves. “The app is a critical response to trends we see in quantified self, big data and surveillance, but we don’t believe any of these things are black and white,” McCarthy says. Which is why she says it was important to build an app that went “beyond speculative fiction design”. It forces us to ask questions and consider the implications of the ownership of personal data, and whether its use could improve our lives.
As University of Canberra sociologist Deborah Lupton writes, the appeal of self-quantification technologies such as pplkpr is that they “render visible elements of one’s self and body that are not otherwise perceptible” – such as emotions – while reminding us that data tells only part of the story. Lupton calls this enmeshment of data and bodies “data doubles” that “have their own social lives and materiality quite apart from the fleshy bodies from which they are developed”.
Pplkpr, however, does not make this distinction, so it’s easy to see it could draw misguided inferences about the value of a relationship. Once, after meeting a good friend, I measured my heart rate and found it slightly more elevated than normal. I was also feeling anxious, and dutifully entered that into pplkpr, but I was certain neither of these things had much to do with my friend’s presence and more to do with other things going on at the time – mounting deadlines and some reservations I had over a public speaking invitation I’d recently accepted.
As part of the app’s testing phase, eight first-year students at Carnegie Mellon wore a heart rate band while using pplkpr for a week. The students were “really open-minded”, McCarthy says, and “they didn’t seem to have as many preconceived biases against the potential for this kind of technology to have a positive impact on their life”.
The algorithm got some things wrong: the app automatically blocked a pair of best friends who appeared to exhibit high levels of stress around each other. “They’re friends again, and have unblocked each other,” McCarthy says. But another student said that “using the app as a justification for not wanting to spend time with someone is a lot more definitive than saying I’m uncomfortable”. That aspect could be by design: McCarthy wonders whether “just like we can chalk up our faux pas to autocomplete or our spam filters, could an app like this provide the excuse or justification we need to say and do what we really feel?”.
Perhaps the allure of pplkpr stems from its simple solutions to the complex “problems” of friendship and human interaction. Correlation does not imply causation – an adage equally applicable to statistics and to our human relationships – but careful interpretation can perhaps provide us with clues. Friendships aren’t always constant, smooth sailing, or born from the purest of motivations. They’re subject to ebbs and flows, and might fulfil a categorical purpose, forged from necessity, obligation or habit. Humans are pretty good at self-sorting, and can determine where a person fits on the scale. But social media lumps everyone together, even though not all friendships are equal. This disconnect is perhaps pplkpr’s most enduring insight, and one we often forget.
The emotional analytics of pplkpr could reveal things about our elective affinities that we are reluctant to admit even to ourselves. But it also proffers a glimpse of what the future of self-quantification might hold: a society where our bodily data takes on a radical new life of its own.