Instagram algorithm provides false alarm for COVID-19, findingsExBulletin

Researchers are concerned that Instagram’s new “suggested posts” feature is contributing to the dissemination of false information.

Dennis Charlett / AFP via Getty Images

Hide captions

Switch captions

Dennis Charlett / AFP via Getty Images

Researchers are concerned that Instagram’s new “suggested posts” feature is contributing to the dissemination of false information.

Dennis Charlett / AFP via Getty Images

According to a new report from a group tracking online false alarms, Instagram recommends false claims about COVID-19, vaccines, and the 2020 US presidential election to those who may be interested in related topics. did.

Imran Ahmed, CEO of the center, said: To counter the digital hatred that conducted the survey.

From September to November 2020, Instagram recommends 104 posts containing incorrect information, or about one post per week, for 15 profiles set up by a UK-based nonprofit organization. Did.

Automated recommendations include a number of photo sharing apps, including the new “Suggested Posts” feature introduced in August 2020 and the “Exploration” section that introduces content that users may be interested in. It was displayed in that place.

This survey is the latest effort to document how the social media platform’s recommendation system contributes. Dissemination of false informationResearchers say it was accelerating last year, boosted by a pandemic and a fierce presidential election.

Facebook, which owns Instagram, has been cracking down more aggressively in recent months.It spread that ban Falsehood about COVID-19 vaccine On its platform of the same name and Instagram in February. However, critics say the company’s automated recommendation system isn’t fully committed to how people are exposed to misinformation. They argue that social network algorithms can send people interested in suspicious claims to the rabbit hole of more extreme content.

Ahmed was particularly concerned about the introduction of “proposed posts” on Instagram last year. This is a feature aimed at allowing users to spend more time on their apps.

Users who see all of their recent posts from accounts they’ve already followed will now see posts from accounts they don’t follow at the bottom of their Instagram feed. The suggestions are based on the content you have already absorbed.

“It’s really powerful to put it on the timeline,” Ahmed said. “Most people won’t notice that the information comes from an account they don’t follow. They think,” These are the ones I chose to follow and trust. ” It is very dangerous because it is. “

The Center for Countering Digital Hate should stop recommending posts “until Instagram can show that it no longer promotes dangerous false information” and exclude posts about COVID-19 or vaccines from recommendations. It states.

To test how Instagram recommendations work, a nonprofit teamed up with youth advocacy group Restless Development to get volunteers to set up 15 new Instagram profiles.

The profile followed different sets of existing accounts on social networks. Those accounts extended to reputable health authorities. For advocates of wellness, alternative medicine, and vaccine repellent.To the far-right militia group and those who advertise those who have lost credibility Canon Conspiracy Theory, Facebook banned in October.

Profiles following wellness influencers and vaccine opponents were posted with false claims about COVID-19 and the more aggressive ones. Anti-vaccine content, The researchers found.

But that wasn’t the only recommendation. According to Ahmed, these profiles were also “election misinformation, identity-based hatred, and conspiracy theories,” including anti-Semitic content.

Second, profiles according to Qanon or the far right account were recommended disinformation about COVIDs and vaccines, even if they also followed trusted medical institutions.

The only profile that was not provided with false information was only recognized health agencies such as the US Centers for Disease Control and Prevention, the World Health Organization, and the Gates Foundation.

This survey doesn’t disclose the number of suggested posts reviewed by profile, so it’s impossible to determine how often Instagram recommends incorrect information.

Facebook spokesman Raki Wane told NPR “Share[s] The goal of reducing the dissemination of false information ”disputed the research methodology.

“This survey is five months old and uses a very small sample size of just 104 posts,” says Wane. “This is in stark contrast to the 12 million harmful false alarms associated with vaccines and COVID-19 that have been removed from Facebook and Instagram since the outbreak of the pandemic.”

According to Facebook, searching for COVID-19 or vaccines in apps such as Instagram will show reliable information from prestigious medical institutions such as WHO, CDC and the United Kingdom National Health Service.

“We are also working on improving Instagram Search to create accounts that make it harder to find vaccines,” Wane said.

Researchers have been tracking for some time the duplication of conspiracy theories and how they appear in social media recommendations.Some anti-vaccine activists have begun Posting Qanon content Last year, a spreader drew attention to the story of an unfounded fraudulent election pivot Post false alarms for vaccines.

“The correlation between these communities is fairly well documented,” said Renée DiResta, who is studying false information at the Stanford Internet Observatory. As early as 2016, the Facebook account she used to track the anti-vaccination movement, Pizzagate Conspiracy, the predecessor of Canon.

Ahmed duplicates the various plots recommended in his group’s study Riot at US Capitol..

“That’s exactly what I saw on January 6th,” he said. “This is a collection of these fringe forces, and what was partly driving it? Algorithms.”

Editor’s Note: Facebook is one of NPR’s financial backers.