In October, Brazil will hold presidential elections that appear to be headed in the same contentious direction as the U.S. presidential election of 2020.
Even before voting has begun, the far-right incumbent, President Jair Bolsonaro, has already started talking about voting irregularities—raising concerns that he might not concede defeat should he lose. Bolsonaro, who is polling behind the leading candidate, Luiz Inácio Lula da Silva, has also questioned President Joe Biden’s election win. “We don’t want that to happen in Brazil,” he said in a recent TV interview.
Disinformation was a key strategy in Bolsonaro’s 2018 presidential win. At that time, voters were barraged with fake information through WhatsApp groups. One analysis found that 56 percent of the most-shared political images on WhatsApp during the election period were misleading. A post-election analysis found that the vast majority of inaccurate information circulating on WhatsApp in the 2018 elections favored Bolsonaro.
This time around, the tech platforms are scrambling to assure the public that they will do better at curbing disinformation. “We learned a lot from 2018 and we are doing a lot more than we did back then,” Dario Durigan, head of public policy for WhatsApp in Brazil, recently told the Financial Times.
Durigan said that WhatsApp has improved the artificial intelligence it uses to identify accounts that are sending out mass numbers of messages and has reduced the number of times users can forward a message from 20 to five.
But experts worry that such measures may not be enough to prevent a Brazilian replay of the “Stop the Steal” movement in the U.S., which culminated in a violent attempt to impede the transfer of power on Jan. 6, 2021.
To understand the Brazilian election landscape, this week I spoke with Patricia Campos Mello, a leading Brazilian journalist who has been tracking the elections and disinformation. Mello is an editor-at-large at Folha de S.Paulo newspaper and an associate research scholar at Columbia University, working on a project on electoral disinformation ahead of the Brazilian 2022 presidential election.
Mello has been a reporter for more than 25 years and has covered international relations, economics, and human rights. She was awarded the Columbia University Maria Moors Cabot award in 2020, the International Press Freedom Award of the Committee to Protect Journalists (CPJ) in 2019, the Vladimir Herzog Special Award for Democracy and Justice in 2019, the International Committee of the Red Cross Prize for humanitarian journalism in 2017, the King of Spain Journalism Prize in 2018, and the Petrobras Prize (2018 and 2017). In 2020, she was awarded the Ordre national du Mérite by French president Emmanuel Macron.
She is also the author of the book “A Máquina do Ódio: Notas de Uma Repórter Sobre Fake News e Violência Digital” (Companhia das Letras, 2020), about populist disinformation campaigns in Brazil, India, and the U.S., as well as threats to freedom of the press in Brazil.
Our conversation, edited for brevity and clarity, is below.
Angwin: In 2018, you published an investigation about election disinformation circulating on WhatsApp in Brazil. Can you talk about what you found?
Campos Mello: I was not a tech reporter. However, I had been covering elections around the world, and in 2008, when I covered the Obama campaign, I started doing stories about big data, and at the time everybody was so full of hope about the potential of big data. Jumping forward to 2014 in India and 2016 in the U.S., there was this epiphany where we realized the harms of these new technologies.
During the 2018 elections in Brazil, there were stories noting the huge number of WhatsApp groups that were sharing political disinformation. These groups just kept growing and proliferating because administrators could add random numbers to them. Meaning you can be added to a group without agreeing to join it. For example, in the last two weeks I was added to six different WhatsApp groups. One of the things I wanted to understand was how these groups were disseminating messages and growing so quickly. Some of it seemed very fast and very organized.
At that time, WhatsApp was claiming that everything was organic and that there was no automation. What I found was that marketing agencies were actually providing mass messaging software that emulates WhatsApp. It was actually very primitive, like a troll farm of people with dashboards sending thousands and thousands of messages to thousands and thousands of groups. People could either buy specific databases of names of voters with specific profiles, or some of the candidates, like Bolsonaro, already had their database. At the time, this mass messaging was not in itself illegal. Rather, it was illegal if it included negative ads about the candidates or if it was not declared in the expenses of the candidate.
Angwin: What did these messages say?
Campos Mello: The messages were either regular political ads or character attacks. One viral message accused one of the candidates of being involved in a corruption scandal. There were also the traditional disinformation ads that claimed voter fraud was happening. It was mostly Bolsonaro supporters who were doing this, but there were also cases of the opposition party engaging in this type of messaging.
There were some studies from the University of Minas Gerais that classified which messages were the most common. They found that posts about voter fraud—that votes were not being counted, or that the voting machines were fraudulent, or broken—were the most common. There were also a series of memes, posts, and videos that became famous at the time saying that the opposing party, the Workers Party, which was running against Bolsonaro, was distributing penis-shaped baby bottles in elementary schools.
Angwin: After this investigation you were subject to a mass harassment campaign. Can you talk about how you dealt with that and if it affected your work?
Campos Mello: In the beginning, they started circulating fake pictures of me with the opposition candidate, saying, “This is the journalist. She’s communist.” There were also fake news and memes with my face saying that I committed a crime. I’m used to working in places that are actually very dangerous for the people who live there, so in the beginning I thought this is just some crazy internet people, they won’t actually do anything. But people started calling my phone and sharing the public events I was going to with email lists of Bolsonaro supporters. It got to the point where I couldn’t go out of my house. Every time I left my building, this neighbor from next door would open the window from the second floor and start yelling “you communist bitch.” On top of all of that, the president himself sued me, demanding that I disclose all the sources for the stories I had already published. I continued to work and to investigate during all this, even though the newspaper hired a bodyguard who had to go with me. I even continued to cover presidential visits.
Then it got worse because one of the sources testified in a congressional hearing in February 2020 and implied that I was trying to seduce him to get exclusive information. I had all the message exchanges and all the audio of the interviews documented, and I just thought, “O.K., fine, I will publish this.”
After he said that, one of the president’s sons, the one who was a deputy, started tweeting and doing videos saying that I was trying to seduce sources to get exclusive information, and he went to the Congress and said the same thing. After he said that, this universe of extreme right-wing people, bloggers, legislators, and ministers, started making jokes. A few days later, the president himself was on live TV, and he made a joke suggesting that I was willing to engage in anal sex in exchange for a scoop against him. He said that on live TV.
So after that, can you imagine the kind of stuff I started getting?
It got to a point where I felt I had to do something, so I sued the president, I sued his son, and I sued other prominent people who had posted and said crude things. I thought that at least it shows this is not normal. They had a history of doing this with journalists, especially women, but before me, it was not as crude and personal. We won the cases against Bolsonaro and his son, but they are both appealing.
Angwin: What action has been taken by WhatsApp and the Brazilian government after your 2018 report. Do you think things will be different in the upcoming election?
Campos Mello: A year later, someone from WhatsApp finally admitted that mass messaging had been used to try and manipulate the results of the election and influence voters. After the election in January, WhatsApp implemented a friction mechanism that reduced the number of messages you could forward to people.
The electoral authorities also banned mass messaging following this. There was also a ruling by the Supreme Court last year saying that, starting in 2022, if you are spreading false information about the elections, which is basically what the president does every day, this can lead to the annulment of your candidacy.
So theoretically, we have an electoral court resolution saying that mass messaging for electoral disinformation purposes is banned and a Supreme Court ruling that sets this precedent, but we don’t know how this is going to be applied in the elections, whether they’re actually going to enforce it and how. Additionally, everything is encrypted, so how are they going to track who’s spreading and sponsoring and paying for the disinformation?
At this point, we have a president that says every day that the elections are going to be rigged. He says he’s going to hire a private auditing company to audit the elections and that he’s going to have a parallel vote counting by the army. It’s like a slow motion coup. When we ask the internet platforms, how are you preparing for this? What are you going to do if you have the president or one of his allies hosting a live video on Facebook saying, “You should go confront poll workers because they are stealing the election”? They don’t seem to have a plan.
Angwin: You mentioned it is like a slow motion coup. Do you think that democracy is in peril?
Campos Mello: I used to think, “Oh my God, these people are so hysterical.” But now, everybody I talk to is really concerned. We essentially have three possible scenarios. One is he wins reelection and with reelection, he gets to pick another two justices for the Supreme Court, so all the checks and balances are gone.
The other scenario is that former president Lula wins, but there is no way President Bolsonaro is going to have a peaceful transition of power. He already has said this. If he says the results are not trustworthy, what’s going to happen? Is he going to be successful in a coup? What kind of social unrest and violence will happen? I think about Jan. 6, and the fact that Brazil is a much younger democracy. I’m really worried. Everybody knows this is going to happen, because every single day he says these things. You have the electoral authority saying every day that this is very dangerous, and all of us are just sitting and watching.
Angwin: What do you think the platforms should be doing to help prevent this slow motion coup?
Campos Mello: First of all, they need to have clear and transparent moderation policies, especially relating to elections everywhere, not just in English-speaking countries. For example, the community standards rules for YouTube were only a translation from the U.S. They mentioned false claims related to citizenship of the candidate, which disqualifies someone from running for office, in a clear reference to what happened to Barack Obama. They banned content that falsely pointed to frauds in elections, but only when it concerned the U.S. or the German presidential elections. After a lot of pressure, they announced a month ago that they would start removing similar content in Brazil. TikTok banned false allegations about mail-in voting—which does not exist in Brazil.
We also don’t know how well their policies are enforced. In Brazil, the platforms received a request from the office of the Federal Prosecutors in São Paulo to disclose how many people they had in the moderation teams speaking Portuguese and how much they were investing in AI in Portuguese. They did not disclose that because they said it was a trade secret. So how do we know if they’re really fighting disinformation? I know it sounds cliché, but we need transparency and accountability.
As always, thanks for reading.
(Additional Hello World research by Eve Zelickson.)