This article was first published here in Portuguese by Agência Pública and translated to English by Matty Rose from the Latin America Bureau.
“Robe-wearing bums, get ready,” a Brazilian man warned in a video made on Kwai, a short-form video platform owned by Chinese social media giant Kuaishou. Kwai is one of the top three most popular social media platforms in Brazil, and a top rival to TikTok, owned by Chinese corporate giant ByteDance.
“Not only are we going to invade the Supreme Federal Court, we’re going to hang you upside down,” the man threatened.
During the campaign, Bolsonaro galvanized supporters by taking aim at the integrity of Brazil’s courts, periodically vowing to appoint new judges to the Supreme Court among other threats to the judiciary’s independence. His supporters on social media have been eager to amplify his anti-judiciary narrative.
“LET’S THROW THE CRIMINALS IN POWER OUT,” wrote another online supporter, referring to court justices. On July 11, this was the second-most shared message in some of Brazil’s most popular WhatsApp groups that support pro-President Jair Bolsonaro.
Nearly two weeks later, Alexandre de Moraes, a Supreme Court judge and current president of Brazil’s Superior Electoral Court, had enough. He ordered the man in the video to be temporarily detained. Moraes is a top target of attacks by Bolsonaro, who has accused the Supreme Federal Court of “interfering” with the exercise of power by the executive branch, and the Superior Electoral Court of “manipulating” the Brazilian electoral system.
Court-ordered detentions of individual bad actors did not put a halt to incendiary online narratives. Misleading and false information circulated on TikTok, WhatsApp, and other social media platforms in the lead-up to Brazil’s national elections on October 2.
Although former president and left-wing candidate Luiz Inácio “Lula” da Silva captured a narrow lead at the polls, he did not secure enough votes to win outright.
That triggered a rematch in a runoff election on this Sunday, October 30. Bolsonaro raised the stakes — he pushed rumors of voter fraud and stated on several occasions that he will only accept the outcome if he wins the runoff.
Digital disinformation has played a pivotal role in Bolsonaro’s rise and now in his fight to stay in power. Agência Pública, Brazil’s first nonprofit investigative newsroom, investigated how short-form videos from platforms like TikTok have helped fuel false narratives and promote pro-Bolsonaro violence.
Using data from the research organization Eleições Sem Fake, or Elections Without Fake News, Agência Pública identified a set of leading pro-Bolsonaro WhatsApp groups and then analyzed the videos that were most widely shared on those groups. The findings showed that nearly half of the most popular videos in these WhatsApp groups had originated on either TikTok or Kwai.
Both TikTok and Kwai have seen explosive growth in recent years. Their interoperability — meaning videos edited on one platform can easily be shared on the other — is part of what makes the platforms vulnerable to the spread of disinformation and hate speech. Tracking the origins of videos is an important tool in countering harmful content, which gained traction across social media in Brazil over the election campaign season.
One of the most popular videos circulated on pro-Bolsonaro social media over three weeks claimed that the 2014 election of former president Dilma Rousseff — who, like Lula, represented the left-wing Workers’ Party — was illegitimate, and sowed doubt about the security of Brazil’s electronic voting system. The two and a half minute clip, edited on TikTok by an account called @fabioprange, reposted part of a documentary made by a Brazilian production company with a streaming platform known as “Netflix of the Right.” The company had shared content that falsely denied reports of deforestation taking place in the Amazon rainforest and distorted Indigenous rights issues in Brazil.
The video was taken down after Agência Pública flagged the video for policy staff at TikTok.
In another video posted by @luizalbertoradio, a man claims that Brazil’s Armed Forces discovered potential cases of fraud in the 2014 and 2018 elections, and that Bolsonaro was going to share evidence confirming this. The channel has nearly 97,600 followers and has received over 790,000 likes for its pro-Bolsonaro content.
The most shared video of the investigation featured former presidents Lula and Dilma congratulating de Moraes, the Supreme Court Justice. The caption on the video reads “Here’s the guy who’s going to take care of the elections. May God have pity on us.”
By early September, the video had garnered more than 883,000 views and had received 23,000 shares, 42,000 likes, and 2,362 comments on TikTok.
Meanwhile, the video sharing platform Kwai has hosted videos claiming the Supreme Court was behind rigged electronic voting machines.
The rampant disinformation trafficked on TikTok and Kwai during this year’s election circulates in a country already overwhelmed with false narratives and concocted news. In 2018, a survey of Facebook and Twitter users found that nine out of ten Bolsonaro voters were exposed to invented, false content during the 2018 presidential elections. This year, according to a study carried out by the Poynter Institute, 44% of Brazilians have said they come across fake news on a daily basis.
Researcher Orestis Papakyriakopoulos, who studies political content on TikTok at Princeton University’s Center for Information Technology Policy, said social media platforms’ lack of transparency freezes study of the problem by researchers.
“Because these [platforms] are black boxes right now, we are talking about things that we usually see as users, but we don’t know how much they happen and we don’t know how exactly the platform contains them and moderate them,” he told Agência Pública.
TikTok’s Brazil press office said to Agência Pública that the company takes “extremely seriously the responsibility that we have to protect the integrity of the platform and the elections.” Kwai stated that it has “security mechanisms that combine artificial intelligence with human analysis to identify and remove content which violates or infringes their policies,” and that it “does not tolerate” content “that has been manipulated with the intention of attack an individual, group, or organization or which attempts to obstruct democratic processes.”
Both platforms have signed partnerships with fact-checking agencies for elections. From April to June of 2022, TikTok reported that it removed nearly 4 million videos for violating the site’s terms. But platform-wide, 0.7% of these were removed for violating rules concerning “integrity and authenticity.” Kwai wrote in its biannual transparency report that between January and June it removed more than 8 million videos for violating its Community Guidelines. It, too, reported that these removals represented less than 1% of the total number of videos posted on the platform during that time period.
