Social media platforms crack down on fake news ahead of Brazil election

Social media platforms say they have launched an effort to crack down on fake news and misinformation in Brazil ahead of elections in October that many expect to be turbulent.

Fake news has proven to be a potent political tool in Latin America’s largest country and was wielded to dramatic effect in the 2018 election of far-right president Jair Bolsonaro.

Caught flat-footed then, Meta-owned platforms — in conjunction with the nation’s electoral court — have since rolled out new technologies to detect and stifle the spread of misinformation as well as streamlined procedures for officials and judges to have content taken offline, according to the company.

Even Telegram, which the Supreme Court in March temporarily ordered to be suspended over its hosting of misinformation, has inked an agreement with election officials to develop tools to flag fake news as well as an AI chatbot to answer questions about the elections.

The messaging app, which allows users to broadcast messages to millions of followers, has been a recent favourite of Bolsonaro, who shifted to the platform after WhatsApp began limiting how widely content can be shared.

“We learned a lot from 2018 and we are doing a lot more than we did back then. We are in a much better place now,” said Dario Durigan, the head of public policy for WhatsApp in Brazil.

“We are making an unprecedented effort. This is the most important presidential election we are having this year by far.”

Stemming the flow of misinformation is particularly important this year as Bolsonaro has gone to lengths to sow doubt about election integrity, notably the country’s voting system. In an echo of former US president Donald Trump, he has repeatedly claimed the country’s electronic ballot boxes are vulnerable to fraud, without offering any evidence.

Many fear the former army captain will not accept the result if he loses, with the head of the Superior Electoral Court warning recently that Brazil could face an event similar to the 2021 storming of the US Capitol.

“Democracy is threatened. Electoral justice is under attack,” said Edson Fachin.

Used by 120mn Brazilians, Whatsapp is a part of daily life in the Latin American nation and played a central role in 2018 elections. But it was also widely abused. Fabricated news and smear stories were spread through the platform, while secretive interest groups backing Bolsonaro bombarded users with mass messages, which have since been deemed illegal by the electoral court.

Durigan said WhatsApp has addressed this by using AI to track suspicious message patterns while immediately blocking accounts that are used to send content en masse. In addition, he said, the app has strived to limit the extent to which messages can go viral by introducing “frictions”.

“In 2018, one user could forward a message to 20 contacts at a time. In 2019 it was reduced to five. In 2020, a message that is forwarded more than five times is labelled as frequently forwarded and that can only be forwarded just once. Now in 2022, any messages that have been forwarded once can only be forwarded to five contacts or one group,” he said.

“Whatsapp bans 8mn accounts worldwide per month. You need to be a human being using it organically, otherwise you are banned,” he added.

The company has also said it would not roll out its new communities function — allowing administrators of chat groups to broadcast messages to thousands of members — in Brazil until after the elections, a decision that Durigan said was made “considering context”.

Sister platforms Facebook and Instagram, meanwhile, say they have begun attaching labels to election-related posts that link users to the electoral court website and fact-checking resources. The groups say they are also running programmes nationwide to train electoral officials on how to get malicious or fake content removed.

“Since 2018 we have made huge improvements and investments in partnerships and in identifying content and making sure we are reducing the reach of negative content. Speeding up our response, this is the main goal,” said Debs Delbart, the programme manager of Meta’s strategic response team.

Delbart added that Meta had also sought to increase the transparency of advertisements by creating a process to verify the identification of buyers and adding disclaimers to show who is responsible for the payment. The information is then made publicly available.

“We are now expanding [beyond politics] the topics where we require this level of transparency, including human rights, the economy and health. If you want to run political or social issues ads in Brazil, you need to be in Brazil. You can’t be [outside the country].”

Critics say, however, the platforms themselves need to be more transparent if they are sincere about tackling Brazil’s fake news epidemic.

“Some of [the developments] are welcome changes, such as more transparency in ad buys, but some of them are either things that were already in place or are only PR moves,” said Patricia Campos Mello, a research scholar at Columbia University.

“For instance, the labelling of election information — we don’t know how well that works and how effective is their moderation in Portuguese, since there’s zero transparency in terms of disclosing how many posts were labelled and how many people did they reach before they were labelled.”

The efforts of the social media platforms have also been partly overshadowed by a dispute over anti-fake news legislation currently before Congress.

Following the rampant spread of misinformation in the 2018 election, lawmakers have attempted to launch sweeping regulations that would, among other things, create strict moderation and transparency requirements as well as a basis to pay news producers for content.

The legislation has been fiercely opposed by Bolsonaro, whose supporters frequently utilise fake news, but also tech groups such as Meta and Google. Meta staff say the bill would “fundamentally change the way the internet works” and would hurt small businesses that advertise online.

Social media researchers, however, say the legislation would require the platforms to disclose more information about their ad targeting and audiences and the tech groups are not comfortable with that.

“The less mandatory moderation, the better for them,” said Campos Mello. “They want to follow their own election and civic rules, instead of following officials rules, because that way they can’t be accountable.”