Instagram’s algorithm pushes users toward content that contains dangerous conspiracy theories, according to a new study.
The report from the Center to Counter Digital Hate (CCDH), a group that tracks misinformation online, found that Instagram is promoting dangerous falsehoods about COVID-19, anti-vax narratives, white supremacist content, and misinformation about the US presidential election. 2020 United States.
The study stated that users were being directed to this content through the Suggested Posts and Explore if they seemed interested in related topics features.
Instagram has dismissed the study, saying it was outdated and used a very small sample size.
Together with a UK-based advocacy group Restless Development, CCDH created 15 profiles on Instagram to follow a variety of trends, pages, and hashtags.
The report indicates that it recommended 104 posts containing misinformation to these profiles in a two-month period from September to November 2020.
“If you follow the health authorities, you never see that the recommendation gives you the wrong things, but if you are looking for wellness or trends that relate to health in general, start giving you anti-vax content,” said Imran Ahmed, executive director of CCDH.
The study focused on a new algorithm launched by Instagram in August 2020, which recommended new content and topics to users in the search bar, to keep them scrolling the platform.
CCDH says that once users have interacted with content about dubious claims, Instagram would suggest more misinformation and extremist content.
“Once you’re hooked on anti-vax content … they were being fed QAnon, white supremacy, electoral disinformation in the United States, really serious stuff,” Ahmed told Euronews.
In a statement to Euronews, the owners of Instagram, Facebook, said that the CCDH study was not representative.
“We share the goal of reducing the spread of misinformation, but this research is five months out of date and uses an extremely small sample size of just 104 posts,” said a Facebook company spokesperson.
“This is in stark contrast to the 12 million harmful vaccine and COVID-19 misinformation that we have removed from Facebook and Instagram since the start of the pandemic.
“We’ve focused on connecting people with credible information, so when people search for COVID-19-related content and vaccines in our apps, we direct them to licensed health organizations.”
Instagram said that the methodology behind the CCDH study was flawed and did not take into account that the platform had proactively identified prohibited content using artificial intelligence.
Between September and November, the company says they removed 1.17 million posts from Facebook and Instagram for violating their harmful disinformation policies.
Last year, the company also accelerated its action on hate speech content targeting people for religious reasons and posts related to the unfounded QAnon conspiracy theory.
“We are also working on improvements to the Instagram search, to make accounts that discourage vaccines more difficult to find,” added the company spokesman Facebook.
But Imran Ahmed rejected Instagram’s statement, reiterating the report’s concerns about the new automated system launched in August.
“When they released an algorithm in the middle of the pandemic, it fed misinformation about vaccines,” Ahmed told Euronews.
“That is an extraordinary failure,” he added.
The CCDH says the 15 Instagram accounts they created are still receiving recommended misinformation on a variety of topics in 2021.
George is Digismak’s reported cum editor with 13 years of experience in Journalism