For months, there were warning signs online that supporters of Jair Bolsonaro, the former Brazilian president, would take to the streets to protest against his left-leaning successor, Luiz Inácio Lula da Silva.
But when far-right rioters stormed Brazil’s key government buildings on January 8, social media companies were again caught flat-footed.
In WhatsApp groups — many with thousands of subscribers — viral videos of the attacks quickly spread like wildfire. Many of the Bolsonaro faithful urged rioters on, calling for a return to military dictatorship, according to encrypted messages reviewed by POLITICO.
On Twitter, social media users posted thousands of images and videos in support of the attacks under the hashtag #manifestacao, or protest. On Facebook, the same hashtag garnered tens of thousands of engagements via likes, shares and comments, mostly in favor of the riots, according to CrowdTangle, the social media analytics tool owned by Meta. This all happened despite Meta pledging to remove any post in praise of the violence.
“They’re not doing enough,” said João Brant, a Brazilian disinformation researcher when asked in October how the social media giants were combating waves of falsehoods. The misinformation was promoted by high-profile politicians and influencers targeting the country’s presidential election battle between Bolsonaro and Lula. Brant is now secretary for digital policies in Brazil’s Social Communication Secretariat, a government agency.
“The very idea of accountability or a commitment — a real commitment — to defend democracy should be part of their responsibilities,” he added. “The idea of no liability at all for the platforms gives them a safe harbor to push the burden to whoever will take the lead in trying to tackle fake news.”
In response, the platforms highlighted efforts taken to quell online misinformation, including: work with outside fact-checkers to debunk falsehoods; disclaimers placed on popular hashtags linked to the Brazilian violence; and commitments to remove content and accounts that glorified the nationwide riots.
Yet in failing to clamp down on such content, the violence in Brazil again highlights the central role social media companies play in the fundamental machinery of 21st century democracy. These firms now provide digital tools like encrypted messaging services used by activists to coordinate offline violence and rely on automated algorithms designed to promote partisan content that can undermine people’s trust in elections.
It also highlights the difficulties in combating long-standing partisan divisions that started well before social media, but have become weaponized by an increasingly sophisticated network of primarily far-right online users — from Brasilia to Berlin to Boston.
In the hours after the riots began across Brazil, for instance, like-minded groups across North America and Europe quickly jumped into action to promote their solidarity with the Bolsonaro supporters and spread those messages worldwide, primarily through Telegram, the encrypted messaging app favored by extremists. That included claiming the Latin American country’s election was “rigged,” akin to allegations promoted by former U.S. President Donald Trump, as well as conspiracy claims that the so-called global deep state was behind Lula’s victory in October, according to scores of social media messages reviewed by POLITICO.
Global far right
“There should be no confusion about the global far-right’s willingness to learn from each other, share tactics, and exploit social media to achieve their ends,” said Wendy Via, president of the Global Project Against Hate and Extremism, a non-profit that tracked Bolsonaro’s use of partisan online tactics during his presidency.
Social media giants did not create the political divisions now engulfing Brazil. But, despite years of promises to slow how such partisanship can spread online, the companies are yet to come to terms with their over-sized role in how democracies function.
In part, it comes down to resources.
Since Elon Musk took over Twitter in late October, the world’s richest man has slashed the internal teams in charge of combating misinformation, including individuals in charge of the company’s oversight in Brazil, according to two people with knowledge of those layoffs, who spoke on the condition of anonymity.
At Meta, the company banned misleading political ads within Brazil, including those questioning the legitimacy of last year’s election. But high-profile politicians like Bolsonaro with large online followings repeated those unsubstantiated claims with little or no censure, while the lion’s share of Meta’s election-protecting resources was earmarked for the U.S. midterm elections in November.
For Damon McCoy, a professor at New York University who has monitored Meta’s response to similar “emergency events,” companies have failed to act quickly enough to delete viral videos, images and partisans news about offline attacks, allowing these falsehoods to circulate widely online.
Instead of focusing on removing posts inciting violence, social media giants should impose a so-called circuit breaker on how their algorithms promote such material, he said. That would limit how posts can go viral until companies’ content moderation teams can respond to real-world threats.
Companies should “push this circuit breaker” to stop offline violence from spreading online within seconds, he said. “You need to have a circuit breaker in the system to realistically handle this kind of crisis event.”