Instagram Launched an Algorithm to Fight Fake News. But Is It Censoring Digital Artists?

Photoshopped images are running afoul of the new fact checkers.

Instagram hit this digitally manipulated image by Ramzy Masri with a fake news label when it was reshared by another user.

Could artists who use Photoshop be targeted by Instagram’s new fake news screeners? The social-media platform rolled out an algorithm to detect fake news last May, but it appears to be getting triggered by digitally manipulated artworks as well.

The alarm was sounded earlier this month by photographer Toby Harriman, who was scrolling through his Instagram feed when he spotted a warning. “False Information,” it read. “Reviewed by independent fact-checkers.”

He clicked on the message and the offending post was revealed: an image by digital artist Ramzy Masri of a man standing atop a hill, Photoshopped with rainbow colors.

Masri first shared the image—part of his #SpectrumEdit series transforming the work of other photographers into otherworldly rainbow-scapes—in February 2017. The original photo, by Christopher Hainey, was taken in California’s Death Valley National Park.

Interesting seeing this pop up for the first time when scrolling the main Instagram feed. Looks like Instagram x…

Posted by Toby Harriman on Friday, January 10, 2020

The warning didn’t appear on Masri’s account, but when it was shared by a third-party account that aggregates images. That account did not identify the photo as a work of fantasy, and thus could have confused potential visitors to the park, leading them in a fruitless search for the perfect multi-colored photo op for their own social media.

The warning isn’t aimed at targeting Photoshopped images in particular, but is meant to keep people from mistaking digital fakery from reality, especially when it’s being shared with the intent to deceive. Posts that have been tagged with the label will not surface on hashtag searches or appear on Instagram’s explore page, and the same warning will be applied whenever it is shared by another user.

“As much as I do love it to help better associate real vs Photoshop,” wrote Harriman, “I also have a huge respect for digital art and don’t want to have to click through barriers to see it.”

The National Coalition Against Censorship also sees cause for concern. “While the goals of this policy may be valuable, it is absolutely likely to do harm to artists,” wrote Nora Pelizzari, the organization’s director of communications, in an email to Artnet News. “Artists manipulate images almost 100 percent of the time, in some way or another—It’s core to what they do. This should not prevent them from posting freely and distributing their work as widely as possible.”

“We don’t hide content because it’s photoshopped, we apply a label when a fact-checker has rated it,” a Facebook representative told Artnet News. “Upon review from the fact checker, they changed the rating, so it is no longer being labelled as false on Instagram and Facebook.”

It is unclear how many artworks have triggered the “False Information” label, but Harriman also found a second example, by digital artist Risvan—a composite of a NASA photograph of the moon and a shot of Horsetail Waterfalls in California’s Yosemite National Park by Instagram user @Markian.b.

“This rating was applied in error and the label has been removed,” said a Facebook representative.

Instagram hit this digitally manipulated image by Risvan with a fake news label when it was reshared by another user.

Instagram hit this digitally manipulated image by Risvan with a fake news label when it was reshared by another user.

The fact-checking website Snopes has existing pages explaining that both images are digitally manipulated—part of an entire section dedicated to debunking possible “fauxtography.” But the boundaries of such a category are hard to define.

“Unfortunately, motivation for altering an image can be hard to objectively determine (say, by algorithm)—is the image altered for artistic reasons or to ‘mislead’ seems both a crucial question and an extremely difficult one to objectively answer,” Pelizzari added.

And even if artists aren’t being directly targeted, this policy could still prevent their work from being shared by accounts with larger audiences, warned Pelizzari. “It would be concerning that the positive publicity and larger audience dissemination achieved by having art press share an artists’ work would be minimized by it coming with a warning label,” she said. The fact remains that artists don’t have control over how their work is disseminated if it goes viral.

Instagram gave the example of a widely disseminated Photoshopped image of a shark on a highway as an example when it announced plans to introduce fact checkers to combat the spread of fake news. Image courtesy of Instagram.

Instagram gave the example of a widely disseminated Photoshopped image of a shark on a highway as an example when it announced plans to introduce fact checkers to combat the spread of fake news. Image courtesy of Instagram.

But given the ease with which misinformation is spread online, one could argue that the benefits outweigh the downsides.

“Yes, these new policies could potentially infringe on artistic liberties, but comparing the tiny impact of the art world against the massive influence of social media in shaping real world politics, this seems like a worthwhile sacrifice,” said digital artist Joshua Citarella. On the other hand, he warned, “this filtration is good optics for Instagram to look like they are ‘fighting fake news’ but leaves other more dangerous propaganda like [influencers paid to post about a music festival in Saudi Arabia] up to an activist user base.”

And, if artists have concerns that digitally manipulated images might be mistaken for fake news, there are steps they can take to avoid inadvertently triggering the algorithms.

“We would encourage them to add context to their images in the descriptions that clearly states this is an artistic work, not a journalistic one,” said Pelizzari. “Artists who feel they’ve been wrongly targeted should first initiate Instagram’s reporting and appeals process.” They can also report the incident to the National Coalition Against Censorship with its online report form.

Article topics