by Joseph Shieber
Imagine a party planner, let’s call him “Mark”. Mark offers his services in party planning to all of the citizens of the town of Happyville.
Now, Mark’s party planning service has an interesting twist.
He begins by throwing a party for a group defined by some close connection — say, a family reunion, or a work party — but then he does something unexpected. He’ll invite some high school friends of the partygoers to the family reunion, or some people with a shared connection to church groups to the work party. What happens then is that people make new connections by getting increased exposure to their friends’ friends — or even to their friends’ friends’ friends.
Almost fifty years ago, the sociologist Mark Granovetter coined the term “weak ties” for these more attenuated connections. There is now a wealth of research tracing the ways that weak ties can affect our lives, from helping us get a job to improving our overall well-being.
Sounds great, right? Mark the party planner is bringing people together, exposing them to new people, and allowing them to harness the strength of their weak ties. What could go wrong?
And let’s allow that Mark really enjoys the way that his party planning service benefits the people he brings together. Of course, he also likes the money that he receives from the advertisers who benefit from the exposure that they get in placing their ads all over his party locations, and the percentage he gets from the businesspeople that he also invites to the party so that they can attract new customers. And he likes it that everyone in town knows him as the guy who makes it possible for them to come together and share news and gossip and just spend time with people who share a connection with them.
But not everything about Mark’s business model is as positive as you might expect! Here are a few ways it might go wrong.
First, suppose that the way that Mark plans these parties has some unintended consequences — consequences that threaten to divide Happyville. Suppose it turns out that Mark’s secret method for linking people through their weak ties also winds up dividing people. Here’s how that might occur.
Let’s give Mark’s secret method for choosing who to invite to a given party a name — let’s call it an “algorithm”. Suppose that the algorithm, good as it is at finding new ways to link people, winds up dividing the town into those who support the high school sports teams on the north side of Happyville — the Red team — and those who support the high school sports teams on the south side of Happyville — the Blue team.
Since the algorithm sorts parties so that the Blue supporters only spend time with the Blue supporters and the Red supporters only spend time with the Red supporters, the two groups soon grow to distrust each other. Eventually, the contests between the two sports teams grow so heated that the school district is forced to cancel all events between the two rivals. The fans’ increasing inability to get along with each other, in other words, was caused inadvertently by the unintended consequences of the algorithm, leading to the destruction of the spirit of fair competition in the town of Happyville. Let’s call this the “Inadvertent Sorting” problem.
Second, imagine that some people have an infectious disease that can spread through physical proximity to others. Maybe at first it’s only a very small group of people within the community — only a few percent of the total population. But now the “strength” of weak ties becomes a great weakness! Mark’s business model of bringing people together allows the infection to spread to many more people than would otherwise be the case and makes quarantining the affected populations much more difficult. Let’s call this the “Infection Multiplier” problem.
The problems, of course, are bad enough in themselves, but they can also combine in ways that make them even worse. “Inadvertent Sorting” and the “Infection Multiplier” could work in concert, so that one population — say, the supporters of the Red team — could be completely overrun with that infectious disease that initially only affected a small portion of the overall population of Happyville.
So far, we’ve only imagined some of the problems that could arise inadvertently — either “Inadvertent Sorting” or “Infection Multiplier” by themselves, or working in concert. But the problems could be even more devastating, if a bad actor set their mind to harnessing the effects of Mark’s business model for nefarious aims.
Imagine that someone from the neighboring city of Cloudytown had an interest in weakening the institutions of Happyville. Suppose that Happyville has two very successful amateur theater companies: the Glad Company and the Delightful Players. Suppose further that the success of Happyville’s theater groups has been a thorn in the side of Cloudytown’s theater group, the Grumpy Group.
Now, the citizens of Grumpytown know how Mark’s secret method, the “algorithm”, can lead to inadvertent sorting. So they decide to investigate which common interests are shared only by supporters of the Glad Company, and which common interests are shared only by supporters of the Delightful Players. The Grumpytown citizens then publicize those interests, and emphasize recent stories and events that encourage the citizens of Happyville to focus on those interests more.
Soon, Mark’s algorithm is sorting parties based on these interests of the citizens of Happyville. Now, as a result, partygoers are socializing only with supporters of one or the other of the theater groups, either the Glad Company or the Delightful Players.
Suppose that the citizens of Grumpytown now introduce some reason for these two distinct groups of partygoers to argue — something that might have nothing to do with their allegiance to one or the other of the theater groups, but something that serves not only to divide them further, but also to make them angry at and suspicious of each other. It gets so bad that neither theater group is able to perform any more, for fear of being disrupted by supporters of the rival theater group.
What we have is a new version of the “Sorting” problem — but no longer inadvertent! This “Sorting” problem was planned — and facilitated by Grumpytown’s ruthless manipulation of Mark’s algorithm. Call this the “Nefarious Sorting” problem.
And like the “Nefarious Sorting” problem, we can also imagine that a particularly unscrupulous person could even introduce an infection into a subpopulation of Happytown, creating a planned version of the “Infection Multiplier” problem — call this the “Nefarious Infection Multiplier” problem.
Given the potential for “Inadvertent Sorting”, “Infection Multiplier”, “Nefarious Sorting”, and “Nefarious Infection Multiplier” problems, I’m not sure that the citizens of Happytown should be content if Mark simply says that his business model just “brings people together” and that any problems resulting from the model are ultimately the responsibility of the people of Happytown, rather than his responsibility. In fact, it’s a tried and true maneuver of destructive industries to try to convince the public that the negative consequences brought about by those industries are not systemic problems but rather the responsibilities of individual citizens or consumers.
This fact already gives us reason to think that Mark should be held responsible for at least some of the potential negative consequences of his service. But certainly, once Mark has specific reasons to believe that problems like “Nefarious Sorting” or the “Infection Multiplier” have actually arisen, it doesn’t seem acceptable for him simply to say that he’s simply providing a service that brings people together. It’s his algorithm! He’s the one who continues to invite infectious people to his parties, even after having evidence of their infectiousness!
In other words, as it stands I’m not sure the citizens of Happyville should take comfort in the fact that Mark has faith that they’ll work it out, because, at the end of the day, he believes in people.