Meta, YouTube failing to crack down on Brazilian election misinformation: report

Meta and YouTube are failing to crack down on posts and ads spreading misinformation about the Brazilian elections despite commitments from the companies to take action against the false narratives, according to a report released Wednesday by SumOfUs. 

Researchers reviewed thousands of social media posts, ads and videos related to Brazil’s elections and found an “ecosystem of content seeking to undermine the electoral process” by spreading false information amplifying doubts about the integrity of that process, according to the report.

It mirrors criticism raised about how the companies failed to crack down on the false “Stop the Steal” election narrative in the U.S. in 2020 and amplifies concerns that tech giants have not improved on their misinformation strategies ahead of Brazil’s October elections or the U.S. midterms in November. 

For Meta, the parent company of Facebook, researchers searched the platform’s ad library using seven keywords related to the elections for content during 2022. For two of the terms, “STF,” referring to the Brazilian Federal Supreme Court, and “Alexandre De Moraes,” a justice of the court, the number of results was too great to be manually processed, according to the report, so researchers reduced the time span to seven days.

The researchers behind the report discovered 56 ads peddling election disinformation that collectively reached 3 million impressions. 

The ads spread baseless claims about electronic ballot boxes, attacked the credibility and integrity of institutions overseeing the elections, and promoted civil unrest in response to an unfavorable election outcome, according to the report. 

The report says that the “high number of video ads in the sample” indicates that Meta’s artificial intelligence moderation systems are failing to identify and remove election disinformation in Portuguese. 

Critics have slammed Meta in the past for not putting in place strong enough content moderation methods for non-English language content. 

A Meta spokesperson defended the company’s policies related to moderating misinformation about the Brazilian election in a statement. 

“While we remove misinformation about voting dates and locations, and recently announced that we will prohibit ads calling into question the legitimacy of current and upcoming elections in Brazil, our policies allow ads that discuss, debate, or advocate for election policies. We don’t believe a private company should limit people’s ability to discuss topics debated in the Brazilian Congress. We’ve prepared extensively for the 2022 election in Brazil and are committed to protecting the integrity of the vote,” the spokesperson said. 

As for YouTube, the report slams the company for serving as a “megaphone for disinformation.”

The report analyzed 10 videos published since June from verified YouTube channels that researchers found mimicked the Stop the Steal narrative by casting doubt on the election process. 

Using Meta’s Crowd Tangle tool, researchers analyzed the journey of the videos and how they spread across wide audiences beyond just the YouTube platform. 

Collectively since they were published, the videos gained 22 million views on YouTube and 2.3 million Facebook interactions, according to the report. 

The Hill reached out to YouTube for comment. 

The companies have touted their election misinformation policies ahead of elections in the U.S. and abroad. 

But the SumOfUs report is not the first to call into question the efficacy of the policies. 

A report released by New York University’s Stern Center for Business and Human Rights found Meta, Twitter, YouTube and TikTok have weak policies on misinformation and have failed to enforce them consistently ahead of the 2022 midterm elections in the U.S.

— Updated at 12:48 p.m.