Predictive Analytics: Resisting the age of algorithm

People can choose to be ruled by algorithmic thinking without running a programme to figure out what’s popular. The fact that we have a specific form of technology that makes it easier to squash risk and creativity is hard to separate from wider trends toward repetition
Representative image
Representative image

NEW YORK: I have few stronger opinions about movie characters than my view that Miranda Priestly, the demanding fashion-magazine boss in “The Devil Wears Prada,” is actually the heroine of the movie.

Not an uncomplicated heroine, certainly not a nice person. But a figure to be celebrated nonetheless: a demanding, uncompromising aesthete whose decisions ripple through the wider culture, whose idiosyncratic taste can affect the palette of the world.

In one of the movie’s famous set pieces, Miranda explains how her assistant’s cheap blue — sorry, cerulean — sweater is actually “a sweater that was selected for you by the people in this room,” the bargain-basement endpoint of a complex aesthetic-commercial process that starts with a single brilliant idea.

I’m no fashionista (to put it mildly), but I still love that scene. So I was struck by Amanda Mull’s recent essay for The Atlantic suggesting its fundamental obsolescence.

The Miranda Priestlys don’t rule fashion anymore, Mull argues. The algorithm does

Her essay starts with a seeming paradox: At a time when we “have more consumer choice than ever, at least going by the sheer volume of available products,” she writes, “much of the clothing that ends up in stores looks uncannily the same.”

The explanation, she suggests, is that fashion is increasingly separated from “the ideas and creative instincts of individuals” and directed instead by a combination of cheaper production models and forecasting systems that takes “the guesswork out of trends.”

The production churns out fashions; the algorithm doubles down on whatever sells the fastest.

Predictably, “when enough brands and retailers begin using these inventory tactics and trend-prediction methods, the results homogenise over time.”

Everything is popular, but nothing is the thing you didn’t know you wanted. And even outfits that look superficially novel are usually repurposings — “gussied up with new details” but the same dress underneath.

This algorithmic repetition isn’t just a fashion trend; it’s the prevailing spirit across multiple cultural domains.

What Mull observes about clothes, the critic Ted Gioia has been analysing in music, where the Spotify era delivers what’s already tested and popular while the opportunities for new artists diminish.

Instead of entering a process of discovery, the online music browser is constantly borne backward — and not into some consciousness-expanding communion with classical-music history but just back to Bruce Springsteen, Paul Simon, David Bowie, an endless boomer-era loop

According to recent market research, Gioia observes, “the new music market is actually shrinking,” even as “the largest investments in music are the acquisition of old publishing catalogues, while almost nothing is spent developing new artists.”

And this tracks with developments in film and television as well — the rule of superheroes, the box-office dominance of aging movie stars and the feel of a certain kind of streaming television, usually on Netflix, that seems to have been scripted by an AI in imitation of 16 other hits

But I don’t want to blame these patterns on technology alone. People can choose to be ruled by algorithmic thinking without running a literal program to figure out what’s popular. And the fact that we have a specific form of technology that makes it easier to squash risk and creativity is hard to separate from wider trends toward sclerosis, repetition, what I spent an entire book calling decadence.

Consider a couple of recent controversies in medicine and medical research, fields distant from Miranda Priestly’s world.

First, there was the revelation that billions of dollars and years of Alzheimer’s research were based on papers that appear to include significant fabrications.

If it pans out, this is a remarkable example of the medical establishment marching into an incredibly expensive blind alley, without skeptics getting a full hearing for about a decade and a half.

Second, there’s the continuing discussion, pegged to a pair of studies that came out this spring and summer, about how and whether the most commonly prescribed antidepressants actually work.

Some of the new research has been over-read by psychiatry’s critics; the assumptions that depression has important chemical components and that antidepressants help people, especially people with severe depression, have not suddenly been overturned.

But both papers add to the strong suspicion that these drugs are oversold and over-prescribed — that we’ve made them a default response to late-modern misery based more on hopeful groupthink than on certain evidence.

Possible fraud and possible over-prescription are different kinds of problems, but they both illustrate how bad cultural and institutional incentives can deaden creativity as surely as Netflix’s algorithms.

A flood of research dollars and prescriptions going in the wrong direction, because everyone wants to imitate everyone else, is the scientific equivalent of everybody making the same dress because that seems to be what the consumer wants — no literal algorithm at work, just a hive mind in which a dissenting voice struggles to be heard.

This kind of system isn’t impermeable to innovation or critique; otherwise, scientific fraud would never be found out, and Netflix wouldn’t have recently lost nearly one million subscribers.

But resisting the rule of the algorithm takes energy and creativity and courage, and the risk for our culture is that our technological skill and our cultural exhaustion are working together, defending decadence and closing off escape.

Visit news.dtnext.in to explore our interactive epaper!

Download the DT Next app for more exciting features!

Click here for iOS

Click here for Android

Related Stories

No stories found.
DT next
www.dtnext.in