We’re living in a strange and wonderful time, where computers are developing intelligence at such an alarming rate that they now know things about us that we might not know about ourselves. For instance, researchers at the University of California, San Diego, are developing an algorithm that can identify whether you might be a hipster, a goth, or a punk based on what you wear on social media.
To do it, they’ve been analyzing photos of eight different subcultures and looking for haircuts, tattoos and jewelry that signify the various groups. Each attribute is then associated with a different subgroup, and various assumptions are then made about what you might be into, based on your subgroup. It’s targeted advertising—very targeted advertising. It’s not much different from what already happens on sites like Facebook, where advertisers crawl your feed to find out what you like and what you’re interested in (which is why so many single gals find themselves confronted with endless online dating ads). The only difference? It’s entirely visual.
For example, if your photos show you wearing a plaid button up, black-framed glasses, and skinny jeans, the algorithm will classify you as hipster, and begin serving you ads that reflect what it believes are hipster interests (i.e. indie bands, hip clothing labels, and, yes, vegetarian food). People wearing ripped clothing and lots of black might get pegged as punk by the algorithm, and get fed ads for punk accessories, punk shows, or punky makeup brands like Urban Decay.
The algorithm isn’t perfect: It’s still only at around 48 percent accuracy. And it is rather stereotypical. Just because we may dress like hipsters doesn’t mean we don’t want to listen to Katy Perry, after all. But researchers are working on perfecting it to make it more useful to advertisers, and to consumers. Because really, wouldn’t you rather see an ad about some cute new clothing label over another annoying dating site in your Facebook feed? We would.