Computational Propaganda and Data Protection in Brazil
In the beginning of 2015, the Journal of Democracy published a series of studies on democratic decline, including two papers by leading political scientists that reached competing conclusions about the overall health of democracy around the world. Larry Diamond famously argued that there was a democratic recession underway. After decades of democratic ascension and consolidation, the rate of democratic failure was again on the rise, reaching a peak of 13% between 2004 and 2013. Steven Levitsky and Lucan Way, on the other hand, called democratic recession a myth. In their view, the evidence of backsliding was rather scarce and, therefore, the scholarly consensus that democracy is in deep trouble was nothing but an illusion. To them, the prevailing story is not one of recession, but instead of democratic resilience.
Fast forward to November of 2016 when Donald J. Trump was elected president of the US, forcing Americans to confront the fragility of their own democratic institutions and challenging the long-held belief of "American exceptionalism". It is hard to tell whether Levitsky still believes that democratic recession is a myth, but it is definitely telling that just a couple of years later he and his coauthor Daniel Ziblatt published the global bestseller How Democracies Die. In their book, Levitsky and Ziblatt offer a compelling description of democratic backsliding across space and time. The book can be read as a cautionary tale or even an exposé of the illiberal rulebook. Nevertheless, the parallels between past and present are what made the book so bone-chilling when it was first published. Readers in countries that experienced democratic breakdown before were left with a disquieting feeling of déjà vu.
At the same time, there was something new about how autocrats were undermining the predicates of democracy. Democracies were no longer not dying the old-fashioned way. There were no tanks on the streets, no coup d'état, no clear break from the preceding liberal regime. Instead, there was a slow and incremental dismemberment of the predicates of democracy. Nancy Bermeo accurately labeled this phenomenon a change in pace. In her words, “Troubled democracies are now more likely to erode rather than to shatter”. Instead of a full-blown breakdown, the period of recession described by Diamond is characterized by democratic erosion, an “incremental, but ultimately still substantial, decay in the three basic predicates of democracy - competitive elections, liberal rights to speech and association, and the rule of law”.
Democratic erosion can be achieved through an array of different means. The autocrat can smear the reputation of journalists and leaders of the opposition to later dismiss any accusation of wrongdoing as “fake news”. The constitutional system of checks and balances can become a preferred target, especially when watchdog institutions exercise their role and try to keep the autocrat at bay. In order to stay in power longer than it is legally permitted, the autocrat may try to leverage his popular support to amend the constitution and extend (or even extinguish) term limits. In other cases, when the autocrat's parliamentary support is strong enough, a new constitution can be adopted to cement a tilted electoral playing field. If evaluated in isolation, some of these changes may seem innocuous and even legitimate, but when stitched together they give rise to what Scheppele calls a “Frankenstate”. Although the formal elements of a liberal and constitutional democracy may stay in place, its substance is significantly damaged (sometimes beyond repair).
But there still is an underexplored side of the story. In our digital age, democratic erosion is technology-driven. Some authors acknowledge that disinformation campaigns on social media (or “fake news”) are somehow connected to the decay of liberal democracies around the globe, but many stop short of making this discussion a core element of how they evaluate and offer solutions to the problems posed by democratic erosion. In other words, the link between technology and democratic erosion is not just a footnote or a curiosity, but rather a central piece of the puzzle that will help us better understand and address this pressing challenge. In this paper, we use Brazil as a case study to assess how technology is being employed to hurt democracy – especially through the use of computational propaganda across social media platforms – and underscore the importance of data protection as a counteraction to this practice.
Throughout the paper, our objective is to show how democratic erosion and technology are intertwined in the digital age and, more specifically, how President Bolsonaro is using social media platforms to spread computational propaganda and entrench his political standing in Brazil. Furthermore, a key aspect of our argument is that data protection is a cornerstone of democracy in the digital realm. We will advance this position by looking at the Cadastro Base do Cidadão in Brazil, a centralized database created by Bolsonaro and designed to host a huge amount of personal data on Brazilian nationals. Our main concern is that, as it currently stands, the Cadastro violates some of the most basic principles and rules of the Lei Geral de Proteção de Dados or LGPD, Brazil’s General Data Protection Law that was approved by Congress in 2018 and, after an unusually long period of vacatio legis, implemented in 2020.
In the next section, we discuss the relationship between social media platforms and democracy. Just like other new technologies, social networks can be used for good or bad purposes, enhancing or hurting democracy depending on the interests of particular political actors. In the third section, we offer a brief description of the state of computational propaganda in Brazil and how the Cadastro can be misused to advance ideological goals by the current or future presidents. In the fourth section, we elaborate on the features of the Brazilian General Data Protection Law and recount the story of how it came to fruition. In the fifth section, we argue that the Cadastro is inconsistent with some of the basic principles of the LGPD. Finally, in the sixth section, we offer some concluding remarks.