The Fake News Machine: How Meta Endangers Brazilian Democracy

September 30, 2022

Meta is the king of the social media world. Not only is the company’s flagship product, Facebook, the world’s largest social networking platform, but two of its other properties — WhatsApp and Instagram — are also in the top five. With this popularity has come significant controversy, especially regarding the role Meta’s platforms play in disseminating political disinformation and “fake news.”

Lying is hardly a new development in politics, but social media companies have created new opportunities for powerful actors to spread lies at an unprecedented scale and speed. In Brazil, Meta’s platforms were used as tools for nationwide disinformation campaigns that helped far-right politician Jair Bolsonaro win the 2018 presidential election. Facing widespread criticism over this and similar campaigns elsewhere, Meta said it would take decisive action to stem the spread of lies. Yet the evidence so far suggests that disinformation continues to run rampant on its Brazilian platforms. 

This time around, the tidal wave of “fake news” circulating on Meta’s social media platforms is helping to promote Bolsonaro’s Trump-like effort to discredit this Sunday’s presidential election, which polls suggest he will ultimately lose. This new, largely unchecked disinformation campaign could have devastating consequences for Brazil, and for the planet. 

Lies, Damned Lies, and Posts

A national law passed in 2017 made a number of changes to Brazilian electoral law, including a reform that allowed candidates to pay to boost content on social media. While this measure moved forward, another controversial proposal to require social media companies to remove “fake news” and hate speech was vetoed by then (interim) president Michel Temer. In other words, while restrictions on online campaigning were loosened, online disinformation remained unchecked.

When the 2018 election season began, online disinformation campaigns became nearly unavoidable. According to a former data scientist at Facebook, the platform’s enforcement efforts often fell far short of its own guidelines – yet this limited effort still identified and removed “10.5 million fake reactions and fans from high-profile politicians in Brazil and the US in the 2018 elections …” Facebook campaigns were only the tip of the iceberg; Meta’s other platforms would pose an even larger problem.

In a scandal that broke after the first round of Brazil’s 2018 elections, it was revealed that wealthy businessmen had illegally funded an operation obtaining millions of phone numbers from Facebook and sending political disinformation to their WhatsApp accounts. The app is widely popular in Brazil as a free alternative to text messaging, and 48 percent of online Brazilians rely on it as a source of news.

The scale of the WhatsApp campaign was enormous. One of the false news stories it promoted — that Bolsonaro’s opponent was sending out “gay kits” to schools in order to indoctrinate children into homosexuality — had a far-reaching impact in Brazil, much of which remains deeply conservative on social issues. This blatantly fake story was seen by 74 percent of voters, and 56 percent of those who saw it say they believed it. According to the same survey, a shocking 93 percent of Bolsonaro voters reported that they had seen stories (falsely) claiming that ballot boxes were rigged. An analysis of viral WhatsApp messages sent during the campaign found that messages spread by right-wing sources were 14 times as likely to contain verifiably false information as those from left-wing sources.

Democracy on the Brink

The sheer size of the WhatsApp operation during the presidential campaign provoked serious concerns about social media abuse by the newly elected Bolsonaro — whose online presence has been controversial to say the least — and his supporters. Four years later, the stakes are even higher. 

This time, Bolsonaro has struggled to break out of a distant second place in polling, and is now under judicial investigation for repeated attempts to discredit the Brazilian electoral system. He has repeatedly suggested that he may refuse to accept the election results if he loses, raising fears of a coup or a sustained destabilization campaign that could severely damage the country’s democratic institutions. If he wins, he has promised to continue to further pursue policies that are devastating the Amazon rainforest and Brazil’s indigenous communities.

As tensions rise ahead of the October 2 first-round election, so has the level of political violence. Researchers have found that the rate of attacks against Brazilian politicians has quadrupled since Bolsonaro took office, with 214 attacks in the first half of 2022 alone. In recent weeks, Bolsonaro’s primary election challenger — former president Lula da Silva (commonly referred to simply as “Lula”) — has had a campaign event bombed and two supporters murdered by followers of Bolsonaro. One of the killers posted frequently in support of Bolsonaro on Twitter and even posted a photo of himself meeting one of Bolsonaro’s sons (who are accused of aiding the disinformation campaigns supporting their father).

Meta has a history of standing by while its platforms are used to promote violence, including bloody ethnic cleansing in Ethiopia and Myanmar. Yet the company claims to have since improved its moderation and enforcement policies. An August 12 statement from Meta on the Brazilian election asserts that “In recent years, we’ve increased our efforts to combat misinformation …” Meta now considers Brazil to be a top priority, but this still doesn’t mean that it is actually enforcing its own rules. 

But for all of the company’s claims of improvement, anti-disinformation activists report that Meta’s platforms are still flooded with both “fake news” and extremist content. A Facebook ad incorrectly claiming that Lula is linked to Brazilian organized crime has been viewed at least 200,000 times. WhatsApp memes call for “[m]ilitary intervention with Bolsonaro in power now!” and declare “[w]e are ready for war!”

Human rights organization Global Witness decided to test Facebook’s moderation in Brazil by submitting 10 advertisements designed to break the platform’s rules, “five containing false election information and five aiming to delegitimise the electoral process.” They did not verify themselves and did not include a “paid for by” disclosure in the ads, both of which are required according to Facebook rules. Nonetheless, Facebook approved all 10 of the ads.

Meta platforms can also be used to dodge some of the restrictions that Brazilian law places on campaigning. Electoral rules in Brazil regulate when candidates are allowed to officially start campaigning, banning any official political advertisements before August 15 of this year. But these rules don’t prevent groups outside of the campaign from paying for ads that support their candidates. Thus, there have been at least 20 instances of Facebook and Instagram approving outside ads for display before the campaign’s legal starting date. Among these advertisements, pro-Bolsonaro content received more than 16 times as many views as pro-Lula content.

The damage that all of these abuses can cause is amplified by Facebook’s explicit policy of allowing lies in political advertisements, a decision that has been criticized by hundreds of the company’s employees. Bolsonaro tried to further weaken these already loose standards by attempting to make it illegal for social media platforms to remove posts for disinformation; the move that was quickly struck down by the Brazilian Supreme Court. The following month, Facebook and Youtube removed a video of him claiming that vaccines cause AIDS.

Lies for Cash

Brazilians have been working overtime since 2018 to find new ways to address this flood of electoral manipulation. Electoral authorities have created a “Counter Disinformation Program,” and campaign laws have been updated in an attempt to outlaw “the dissemination or sharing of facts known to be untrue or seriously decontextualized, damaging the integrity of the electoral process.”

Until Meta’s platform policies change, however, legal attempts to mitigate these issues are unlikely to be effective. For this reason, a coalition of over 90 civil society organizations in Brazil is proposing reforms to combat distortions and fabrications online without infringing on freedom of speech. These include stronger rules against electoral disinformation, greater transparency about platform standards, updated terms of service regarding political violence, and a right to appeal for anyone who has their content removed.

Central to this issue is that platforms like Facebook make money from high user engagement, which attracts advertisers and gives the company personal user data to sell. This means that social media companies have a profit incentive to accept any content that results in high engagement, regardless of its truthfulness or even its legality. Yael Eisenstat, the former head of election integrity at Facebook, agrees: “Facebook profits partly by amplifying lies and selling dangerous targeting tools that allow political operatives to engage in a new level of information warfare.”

With just days left before Brazilians head to the polls to decide the political future of their country, the use of social media by political elites to manipulate elections is again in full swing. Democracies are facing unprecedented threats around the world, and Brazil illustrates the dangerous effects that for-profit social media companies can have on our access to information. Until they are forced to reform, these companies will continue to exchange lies for cash.

Support Cepr

APOYAR A CEPR

If you value CEPR's work, support us by making a financial contribution.

Si valora el trabajo de CEPR, apóyenos haciendo una contribución financiera.

Donate Apóyanos

Keep up with our latest news