Tuesday , September 22 2020
Home / ARTIFICIAL INTELLIGENCE / By fighting “fake news”, does Instagram censor memes and works of art?

By fighting “fake news”, does Instagram censor memes and works of art?

Graphic designers, authors of ultra-retouched images, have seen their work labeled “false information” by Instagram.

Creative images taken for “fake news”? Instagram launched a new anti-info tool in October 2019. In theory, it's a way to limit the spread of fake news. In practice, artists pay the price. This is the case of the publication below, which shows an unreal landscape of mountains in the colors of the legendary rainbow: “Wouldn't you prefer to be here?” It is the work of a graphic designer, an artist, therefore, whose trademark is to give landscapes and objects the vibrant colors of the rainbow. However, here it is adorned with a mention “False information”.

How does this anti-fake news tool work? Instagram (owned by Facebook) has partnered with theInternational Fact-Checking Network, a network of around 70 independent fact-checkers. Our section “True or fake?” is also one of them. When one of these verifiers detects false or partially false information, Instagram receives a notification and applies two measures.

The image is first removed from the “Search and Explore” and “Hashtags” sections to reduce its distribution. It remains accessible only by consulting the account of the user who shared it. Then, a mention “false information” bars the publication and a link to the article explaining why the image is false is added. Instagram tracks identical publications, using an algorithm, to affix the same marking as many times as necessary.

Users can also report information themselves, thanks to the general reporting tool, which now provides the reason “false information”, alongside cases of “hate speech or symbols” or “nudity or sexual acts”. “.

So how did this rainbow mountain photograph get targeted by anti-info radar? At first, neither the photograph nor the legend caught the eye of Instagram. It all started with an article from the Indian media NewsMobile. Member of the fact-checking network with which Instagram is associated, the media spotted the original photo in natural colors and on the other hand reported several shares of the rainbow mountains with the only caption “Rainbow mountains, California “or” Death Valley National Park “. So simple color retouching means that a publication is classified as false information?

“We do not hide content because it is retouched, we apply a label [“Fausses informations”] when a fact-checker has evaluated it, defended a spokesperson for Facebook, to Hypebeast, January 16. From the fact-checker's examination, [Instagram] has modified [la classification de cette photo de montagnes arcs-en-ciel], it is therefore no longer categorized as false on Instagram and Facebook.“Indeed, the publication is no longer labeled” fake news “.

These colorful mountains are not an isolated case. One example among others: this bright and orange moon that pours into a waterfall is not fake news, and yet it bears this label. In question, there too, an article from the American site Snopes, which details the strings of this ultra-retouched image, composed of at least two other photos. The image is actually the work of a digital artist and not of a news broadcaster.

A photo of the Moon pouring into a waterfall considered on Instagram to be fake news. (INSTAGRAM SCREENSHOT)

Memes (these viral humorous images) also suffer from this anti-infox policy. This is the case with this montage in which the famous pyramids of Giza are all surmounted by a very bright planet. The image is accompanied by a humorous legend: “Give a note to my pyramids”. At the origin of this meme, a tweet, itself viral, affirming that it is a perfect alignment of the planets Mercury, Venus and Saturn which “occurs every 2373 years”. This inaccurate claim has been verified and denied by several fact-checking sites, such as Snopes (in English) and Maldita (in Spanish). Result: the photo, even presented in another context, as the object of a joke, is considered by Instagram as “fake news”.

Screenshot of the Worsthistorymemes account, on Instagram, January 22, 2020.
Screenshot of the Worsthistorymemes account, on Instagram, January 22, 2020. (INSTAGRAM)

On the other hand, unlike memes and artistic photomontages, the publications of political leaders are not affected by this anti-info tool. The latter was however developed in part to guarantee the smooth running of the 2020 American elections. “We consider that by limiting political speeches, we would also limit access to information from the public, who would be less aware of the words of elected officials”, justifies Facebook, in a press release.

Leave a Reply

Your email address will not be published. Required fields are marked *