Brazil election: Platforms are not curbing disinformation

João Brant is director of Instituto Cultura e Democracia, and coordinates the project Desinformante.

Brazil is a few days away from the first round of elections that will elect the nation’s President, all 27 state governors, federal and state representatives. Four years after an election marked by the intense use of fake news, there was great expectation to understand what the 2022 scenario would be.

The partial assessment points to a significant amount of false and misleading information in circulation, but to a lesser impact on the voting decision. The definitive assessment, however, can only be made after the elections.

What can already be said, though, is that the efforts of digital platforms to contain disinformation have not been enough to prevent the circulation of false and misleading content. If in 2018 the biggest problem was misinformation against candidates, this year a new tier was added with the problem of misinformation against the electoral system itself.

Monitoring from recent years shows an increase, with a 2021 peak, in the circulation of content that brings unfounded allegations of electoral fraud and that aims, through baseless conclusions, to erode trust in the Brazilian electoral system.

At the risk of Brazil repeating, in a worse version, the Jan. 6 U.S. Capitol riot, a series of actors began to demand stronger commitments from the platforms with the protection of electoral integrity. The Superior Electoral Court (TSE), which works on the definition and application of rules on elections in Brazil, established memoranda of understanding with all digital platforms to agree on commitments to combat disinformation. In addition to these memoranda, the TSE has held periodic meetings with the platforms to move beyond the commitment of the agreements.

Last July, more than 100 entities released a document aiming to point out what platforms should guarantee in their policies and in their actions to protect the integrity of Brazilian elections. In addition to pointing out specific recommendations on electoral integrity and on the transparency of political advertisements, the initiative also highlighted necessary measures to combat political violence and against disinformation related to the climate crisis and to the Amazon rainforest.

After two months of dialogue with the platforms, the organisations published an assessment document recently, which has the following conclusions: in general, the platforms have policies to combat disinformation against the integrity of the electoral process, but disinformation against candidates remains with few restrictions, either because of the absence of specific policies - as in the case of Twitter and YouTube - or because of exceptions given to politicians and candidates – as in the case of Facebook and Instagram.

It said that Meta platforms do not have policies that determine action in the face of demonstrably false content that alleges electoral fraud, and that except for Twitter, no platform has a policy to prevent calls to civil unrest against the democratic order or interference in the peaceful transmission of power that do not explicitly call for violence. Telegram remains without effective commitments to action in the face of disinformation and attacks on democracy. The platform has no published policy on the subject.

The report said that all digital platforms report having protocols to react immediately to political crisis situations, but the procedures are not publicised, and there is no clarity on how such companies will assess the seriousness of the acts, and what type of actions will be taken. It also added that the memoranda of understanding signed with the TSE are limited and insufficient to limit disinformation in the electoral process. Nevertheless, its implementation has advanced with platforms except for Telegram, which has not incorporated the promises made into its policies.

Writing days before the first round, it is possible to say that the impact of fake news in the presidential election seems to be smaller than in 2018, largely because society is more prepared to deal with the phenomenon, and this election is between two candidates that are, or have already been, presidents, so are quite well known and less prone to distrust.

However, two main concerns remain: the impact of disinformation targeting specific audiences, such as evangelical voters in closed information environments, and the risk of disinformation about electoral results sparking civil unrest. If these will mean real impact or are just overstated concerns remains to be seen.