Does Instagram encourage clichés about nudity? - France 24

Does Instagram favor a canon of beauty and nudity? A study raises it based on the analysis of photos that appear in the thread of voluntary users, while the censored "influencers" complain about the "errors" of the algorithm.

To be successful on Instagram, nothing beats a lewd pose in a swimsuit or bikini, as can be seen from how common these images are on the app.

American reality star Kylie Jenner appears to have recently applied this precept by uploading a bikini photo to call on her 197 million subscribers to register to vote ahead of the upcoming US presidential election.

A question arises: does the algorithm of the social network, that is, the secret formula that governs the content of the platform, encourage this type of initiative?

According to research from the AlgorithmWatch organization published in June, the answer is yes.

"Our results allow us to affirm that a photo of a woman in underwear or a swimsuit is shown 1.6 times more than one of her dressed. For a man the rate is 1.3", they declared to the French digital newspaper Mediapart two study authors, Nicolas Kayser-Bril and Judith Duportail.

To obtain this result, they analyzed the 1,737 "entries" (posts) of 37 Instagram accounts, followed by 26 volunteers who had installed an add-on in their browsers to count the number of times each image appears.

In 2016, Instagram, which is about to turn 10 years old, stopped presenting photos in chronological order. It is an algorithm that selects them to match the preferences of each user, according to a series of parameters that are not clear.

According to the study authors, it could be based on a "nudity level" calculated when the image is posted online. They cite a 2011 patent filed by Facebook (which bought Instagram the following year), which protects a system for identifying skin through bands of specific colors.

This study is "completely biased," an Instagram spokeswoman told AFP. "The algorithm analyzes the time spent on a certain type of content, the interactions, and shows as a priority" the content that attracts each subscriber, but "there is no patent that gives priority to nudity, it does not make sense."

The sensation of similar images would then be due to the user's habits, who could avoid it by taking "a small step, to go looking for other types of images", explains the spokeswoman.

-Prudishness-

Social networks are accused of reproducing social biases by personalizing the content they offer to users as much as possible. But it's hard to back up these observations because studies often run into a lack of data provided by the platforms.

The case is notorious on Instagram because the application has an economic responsibility with the "influencers" (paid by brands according to their audience), but also with society by transmitting a certain physical standard to more than a billion users users.

At the same time, Instagram is accused of prudishness, and above all of lack of objectivity in the application of its own rules on nudity.

These prohibit "close-ups of fully exposed buttocks" and "women's nipples," but in several cases moderation has been removed to show photos of nude women, showing off their shapes and love handles.

At the beginning of the year, the social network removed images of Internet users holding the cover of the French magazine Télérama on fatphobia.

"The algorithms of Facebook, Instagram and others don't like nudity, even when there is nothing pornographic about it (...) The photo of (the DJ on the cover) Leslie Barbara Butch shows neither sex nor nipples , but obviously a lot of skin. Too much, apparently, for social media," the magazine wrote.

This summer, French "influencer" Juliette Katz, known online for her Coucoulesgirls account that campaigns against beauty stereotypes, complained about censorship: "Is the little brown tip of my nipple considered a ' sexual act'? My skin too present or the fact that my body takes up too much space in the photo?"

Instagram denies any "censorship of a certain category of people. Sometimes we make mistakes, either with the algorithm or humans," the spokesperson acknowledged. But we don't calculate a "skin percentage" to apply the moderation criteria, "that's an urban legend."

© 2020 AFP