This article published on 15 February 2022 was written by Tow Fellow Patricia Campos Mello for Folha De S.Paulo and was translated from Portuguese to English for the Tow Center. It is reposted here with permission and thanks.
The agreements released by the TSE (Brazilian Superior Electoral Court) and internet platforms on Tuesday (15 February 2022) fall far short of the electoral policies adopted by companies in the United States and do not respond to how they will react if the result of the 2022 election is contested and there is incitement to violence, which is the main concern of authorities and experts.
In Brazil, except for Twitter, none of the companies —Google/YouTube, Facebook, TikTok, and Kwai—currently specify how they will react in the event of a massive campaign of electoral disinformation, such as the one that occurred in the U.S. and culminated in the invasion of the Capitol on January 6, 2021, which left five dead.
In addition, the country has President Jair Bolsonaro (PL) and his allies as one of the main sources of dissemination of false statements about alleged fraud in elections.
Twitter updated its policies late last year and lists several fake election contents that will be subject to removal, labeling, or reduction of reach by the platform.
Among them, are “controversial allegations that may call into question faith in the act (voting) itself, such as unverified information about electoral fraud, votes adulteration, counting votes or certifying the results of the election.”
It also cites “misleading claims about the results” such as “celebrating victory before the results of the election have been certified, inciting illegal conduct to prevent the practical or procedural implementation of election results.”
However, YouTube, which belongs to Google, says its policy bans content that “exposes false claims of widespread fraud.” But that applies only to “all past U.S. presidential elections” and the “German federal elections of 2021.”
More generally, YouTube presents a list of banned content that is just a translation of the platform’s civic rules in the U.S., such as false claims about candidates’ qualifications (inspired by the fake news campaign of former President Barack Obama’s home country).
Google, in the memorandum of understanding with the TSE, undertakes to begin publication of the “Political Ads Transparency report”, for ads related to holders and candidates for elective office at the federal level.
However, the understanding of TSE was that the company would release this report in November 2021 —now it says it will be implemented in the first half of 2022.
This political ad transparency page already works in countries such as the UK, the European Union, India, Israel, and the US.
In a statement, Google said it initiated advertiser verification in November 2021 and that the information will be the basis for the report, which will be published in the first semester of 2022.
In the US, YouTube was heavily criticized for acting late in relation to the conspiratorial videos and for not adopting more specific policies for the 2020 election, as done by other platforms.
It was only on December 12, 2020, more than a month after the vote, that the company announced that it would now remove videos with allegations of widespread voter fraud.
Among the measures to combat electoral misinformation promised by Google in Brazil is a Doodle Elections 2022 – “fun, surprising and often spontaneous versions of the Google logo to celebrate holidays”.
According to the company, Doodle will be at the top of searches dedicated to elections, “an efficient resource to reach the millions of Brazilians who search Google every day and increase engagement on the importance of voting and democracy in the country.”
In Brazil, Google announced between its commitments with the TSE the highlight in the Google Play Store applications with civic content, including the Electoral Justice Officials.
But it doesn’t say if there will be any kind of oversight in candidate applications. The Bolsonaro TV app was downloaded more than 100,000 times in the Google store, and the PT, more than 50,000.
More specifically, Google announced an Election Trends Hub, a page with information on Google Search trends, and a reporting channel.
Facebook since 2018 reveals who pays and what is the engagement of political advertising and keeps a library of election ads in Brazil.
However, in Brazil, the library is more restricted than in other countries.
In nations such as the US, Chile, EU, India, Mexico, Myanmar, and the United Kingdom, this transparency of advertisers is also required by buying ads on “social issues”, which cover topics such as civil and social rights, crime, education, environmental policy, weapons, health, and immigration.
In the US, the platform adopted very specific moderation rules for the 2020 election.
In September of that year, it announced that it would put informative labels on content “that delegitimizes election results or questions the legitimacy of voting methods, stating that the process will lead to fraud.”
In Brazil, Meta, owner of Facebook and Instagram, announced that users would see, from December 10 last year, a label on posts about elections in general and would be directed to the Electoral Justice page.
But, according to FGV’s mapping, there are contents posted after that date with false information about the elections and they did not receive labels. In Brazil, the platform has no specific policy regarding content that questions electoral integrity.
In the U.S Meta apps stopped recommending users to join “civic” groups, with some political connotation and restricted the number of invitations that could be sent per day. It also banned content that calls for oversight of polling stations using militarized language.
Facebook and Instagram banned political ads two weeks before the U.S. election and only resumed in March 2021. The company has not committed to doing the same in Brazil.
Twitter banned political ads globally in 2019. According to the company’s founder, Jack Dorsey, “the reach of political messages should be conquered, not bought.”
In the U.S. campaign, the company began removing tweets that were inciting the contest of the election results and to put alerts on uninformative tweets from political figures and profiles with more than 100,000 followers.
A sensitive point, both in the U.S. and in the rest of the world, is Meta’s policy towards politicians. Content published by politicians and candidates does not go through fact-checking checkers.
That is, false information disseminated by politicians has no reduced scope. According to the company, the approach “is based on Facebook’s fundamental belief in freedom of expression.”
In Brazil, Meta has committed to offering the tool called Megaphone for the TSE to disseminate messages, which has already been offered in 2020, and to put the electoral labels that have been already in use since December.
It also announced a chatbot on Instagram for access to information about the election and strengthened a complaints channel for the TSE.
In the case of messaging applications, as content is encrypted, moderation policies are not discussed, but rather ways to reduce the viralization of electoral misinformation.
Telegram, with about 50 million users in Brazil, is not cooperating with the TSE. It has groups of 200,000 members and channels with unlimited numbers. Bolsonaro has almost 1.1 million followers. The company does not respond to questions from the Brazilian justice system, nor does it take measures against disinformation.
WhatsApp, although not explicit in the agreement, committed to the TSE to not launch before the elections the new product “communities”, which would increase the number of users in groups.
The measure went against the app’s efforts to prevent message viralization around the world. To do this, the company has limited the number of referrals for a message from 20 to five.
And subsequently, it’s only a referral, when it comes to viral messages. The company also went on to sue marketing agencies that offer mass distribution, which was banned by the TSE in the 2019 resolution, and created a reporting channel for this.
In the memo, WhatsApp announced an official TSE channel to communicate with Brazilian voters. The measure goes beyond what was adopted by the app in the US in 2020, where there is only one chatbot for which users can write for election information.
Twitter has committed to the TSE to launch in Brazil a system of warnings when voters seek issues related to elections. One to clarify false narratives about the electronic ballot box and the electoral process.
The platform also announced a channel for faster analysis of complaints by the TSE but does not say how it will account for the resolution of complaints.
On January 17, Twitter implemented a feature that allows users to report uninformative content. But the agency Lupa showed that the company still delays or does not remove denounced tweets.
TikTok’s terms of use in Brazil on electoral integrity are a literal translation of American rules, without any adaptation to Brazilian reality, including talking about “allegations of electoral fraud resulting from voting by mail” —something that does not exist in Brazil.
Kwai (which does not have much performance in the USA) does not directly refer to the electoral context in the rules.
But it is the only company that promises the TSE to provide “the rules and policies of civic integrity applied to elections, in the simple, clear and national language, through public URL” to ensure transparency of the platform’s performance in cases of electoral disinformation —which is the main demand of the authorities.