Facebook’s failure to curb the massive disinformation campaign that preceded Jair Bolsonaro’s election victory shows that it is unable to control the dangerous consequences of its own dominance.
The Brazilian elections were supposed to be the event that finally turned the tide in Facebook’s ongoing battle against disinformation. Following two years of intense scrutiny and eager not to repeat the mistakes of 2016, the company clearly wanted to position Brazil’s presidential election as a showcase for its vigilance and improved abilities to combat the spread of fake news and propaganda. It invited reporters into its “war room,” banned hundreds of thousands of accounts, and touted the efficiency of its own efforts.
The election of Jair Bolsonaro and the massive fake news campaign that preceded it, however, showed something quite different. Instead of being a story about Facebook’s triumph, Brazil now joins Myanmar, India, the Philippines, the US and Brexit as yet another indication that Facebook is either unable or unwilling to curb the weaponization of its platforms by malicious political actors.
It also served as a stark reminder why far-right authoritarians love Facebook so much to begin with. In March, after Italy’s far-right Northern League won 17 percent of the vote in Italy’s general election and became the country’s de facto kingmaker, party leader Matteo Salvini felt it necessary to credit Facebook for Lega‘s stunning electoral achievement. “Thank God for the internet, thank God for social media, thank God for Facebook,” said Salvini, a social media savant who has made effective use of Facebook to amplify his party’s anti-immigrant agenda, turning himself into Italy’s most popular politician in the process.
Bolsonaro, a far-right extremist who favors torture and extrajudicial killings and regularly expresses nostalgia for Brazil’s military dictatorship, didn’t immediately praise Facebook following his election victory on Sunday. He did, however, take to Facebook Live for his first post-campaign statement, in which, among other things, he confirmed that he intends to follow up on his campaign promise to purge his political opponents from the left. (Bolsonaro later addressed TV networks with a second victory speech that was markedly different).
Bolsonaro’s choice to make his first address as president-elect via Facebook, instead of opting for the traditional press conference, was telling. Much like other far-right authoritarians, Bolsonaro’s campaign, which began as little more than a curiosity, relied heavily on social media, completely bypassing traditional media outlets (among his favorite targets). Much of this activity took place on Facebook, including near-daily Facebook Live videos. But the real action took place on WhatsApp, which is owned by Facebook and is Brazil’s most dominant social platform and messaging service, with a penetration rate of close to 100 percent of internet users.
WhatsApp is where over 40 percent of voters in Brazil—including 60 percent of Bolsonaro’s voters—get their news from. It is also where, in the run-up to the election, supporters of Bolsonaro inundated voters with hundreds of millions of WhatsApp messages containing false news items and conspiracy theories about his rival, Fernando Haddad. Supporters used scraping software to collect people’s phone numbers and, according to The Guardian, overseas numbers to bypass WhatsApp’s spam-detection tools and send messages in bulk, resulting in a “tsunami” of fake news. Bolsonaro himself professed ignorance, despite reports that the effort was bankrolled by business groups that supported him. (Bolsonaro enjoyed the support of the country’s business elite and reportedly won 97 percent of the country’s richest cities).
Since WhatsApp is an end-to-end encrypted service, it is difficult to measure the scope and reach of the fake news campaign. But according to one study that examined more than 100,000 images that spread through WhatsApp groups focused on Brazilian politics, among the 50 images that were most widely shared, only 8 percent were truthful, and over half were either misleading or completely false.
The problem was compounded by the fact that many in Brazil consider WhatsApp be more trustworthy than traditional news outlets. It certainly didn’t help that newspapers found it difficult to share their content through the app. In a recent piece (translated from Portuguese by the Global Editors Network), Brazilian journalist Pedro Doria, the former executive editor of O Globo, wrote that WhatsApp blocked some newspapers from sending news articles to their readers, considering the messages to be spam. This, he writes, was “directly connected” to the events that unfolded during the campaign:
“Since the press can’t use WhatsApp to distribute its legitimate content to millions of people, the ones using it to do so are unknown individuals or companies that work in the shadows. It is not possible to spread news articles, but it is more than possible to spread misinformation.”
As big as it was, it is unclear just how much of an effect the disinformation campaign had on the final result. The reasons why Bolsonaro’s neofascist agenda found a receptive audience among Brazilian voters are complex and deep-rooted and probably have more to do with extreme inequality, an inherently corrupt political system, and a skyrocketing murder rate. (You can hear Glenn Greenwald explain the background for Bolsonaro’s victory to Luigi Zingales and Kate Waldock here).
Likewise, there are complex social, political and cultural causes for the rise of Italy’s far-right, and for the rise of other populist authoritarians like President Trump in the US, Narendra Modi in India and Rodrigo Duterte in the Philippines, all of whom benefited from the spread of fake news on social media. While Facebook is much-loved by far-right authoritarians, it is far from the only, or even main, factor behind their rise.
But regardless of its actual impact on the election’s results, Facebook’s failure to curb the onslaught of fake news in Brazil matters, simply because this time it made a larger, more concerted effort—and still came up short.
Unlike previous cases where it failed to respond in time to the spread of misinformation and propaganda on its platforms, Facebook attempted to show that it took its impact on the Brazilian election very seriously. WhatsApp banned more than 100,000 accounts (including Bolsonaro’s own son), sent cease and desist letters to the marketing companies that were reported to be behind the disinformation effort, reduced its forwarding limit in Brazil from 256 users to 20, and launched an ad campaign with tips on how to spot fake news. WhatsApp CEO Chris Daniels took to the pages of Folha de São Paulo to publish an op-ed in which he acknowledged the company’s responsibility to “amplify the good and mitigate the bad.” (Even with the recent limits, cautioned Daniels, “both good and bad” information could still go viral). The company was “delighted to see how efficient we were able to be, from point of detection to point of action,” Facebook’s head of civic engagement, Samidh Chakrabarti, told reporters earlier this month.
Ultimately, however, Facebook proved unable to stop the overwhelming current of false news. Partly, this is due to the flawed nature of its response. As Vice‘s Noah Kulwin reported in January, experts in Brazil have been warning for a long time that WhatsApp is a prime target for political actors interested in hijacking it for propaganda purposes. But when some experts contacted WhatsApp before the election to suggest immediate measures, the company reportedly “responded by saying that there was not enough time to implement the changes.” Then there was Facebook’s odd choice to manage the crisis from its corporate headquarters in California. According to Vice, WhatsApp, which doesn’t have any employees in Brazil, didn’t even send a representative to a meeting of the Superior Electoral Court in Brasilia in which the impact of fake news on the election was discussed—instead, a WhatsApp executive participated via video-conference from her office in Silicon Valley.
But the company’s failure in Brazil is also yet another piece of evidence that, as Siva Vaidhyanathan said in a recent interview with ProMarket, Facebook is simply “too big to manage.” Moderating content generated by 2 billion people living in hundreds of countries, using hundreds of languages across multiple platforms, is a daunting task, one that no other company has ever attempted (and one that is unlikely to be accomplished with two dozen people huddled in a dark room in Menlo Park).
In August, Motherboard’s Jason Koebler and Joseph Cox published a wide-ranging investigation into Facebook’s content-moderation practices. Facebook’s content moderation-related problems, they wrote, come in various forms:
“There are failures of policy, failures of messaging, and failures to predict the darkest impulses of human nature. Compromises are made to accommodate Facebook’s business model. There are technological shortcomings, there are honest mistakes that are endlessly magnified and never forgotten, and there are also bad-faith attacks by sensationalist politicians and partisan media.”
The Brazil case seems to embody many of these failures. One of its main takeaways is that even when Facebook makes a considerable effort to limit the proliferation of conspiracy theories and hate speech, it is still unable to respond effectively. This was also evident in Myanmar, where the military launched a years-long disinformation and incitement campaign targeted against Rohingya Muslims that turned the social network into a “tool for ethnic cleansing.” According to a recent Reuters investigation, despite the measures that it took over the past year to fight anti-Rohingya incitement in Myanmar, Facebook seems to be “losing the war” there as well.
To be sure, further study is needed to accurately determine what role Facebook played in Brazil’s election. Yet the events surrounding Bolsonaro’s victory, and Facebook’s failure to prevent the hijacking of its platforms, demonstrate the company’s inability or unwillingness to mitigate the consequences of its own dominance, which is why it remains the weapon of choice for aspiring autocrats and would-be manipulators of public opinion.
For more on Brazil’s election, check out the following episode of the Capitalisn’t podcast:
Disclaimer: The ProMarket blog is dedicated to discussing how competition tends to be subverted by special interests. The posts represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty. For more information, please visit ProMarket Blog Policy.