NetLab UFRJ Director Takes Part in Public Hearing at the AGU to Discuss Meta’s Moderation Changes

Public Hearing at the Attorney General’s Office (AGU): Speech by Marie Santini / Reproduction

This Wednesday (01/22), the Director of the Laboratory for Internet and Social Media Studies at the Federal University of Rio de Janeiro (NetLab UFRJ), Marie Santini, took part in a public hearing called by the Office of the Attorney General of the Union (AGU) to discuss the effects of new content moderation policies implemented by digital platforms in Brazil.

Watch the full video of Marie Santini’s speech at the hearing below, as well as its transcript.

https://www.youtube.com/watch?v=uLYX5Nl2kjY

OFFICE OF THE ATTORNEY GENERAL OF THE UNION

PUBLIC HEARING, JANUARY 22, 2025

PROFESSOR R. MARIE SANTINI

Good afternoon to everyone present. I would like to sincerely thank you for the opportunity to appear before this public hearing to discuss a matter of extreme relevance to Brazilian society: the recent changes to Meta’s policies. The decisions to dismantle fact-checking programs and relax standards for moderating disinformation and hate speech represent a turning point that threatens informational integrity and freedom of expression in Brazil and worldwide.

However, I would like to draw attention to a very significant change announced by Mr. Zuckerberg: alterations to the company’s algorithms, which decide which voices will be amplified or silenced, which content will be recommended, and how advertisements will be targeted to each of us. These algorithms, programmed for content curation and moderation, operate without any transparency over their criteria. We do not know which content is effectively moderated (removed by the company’s own initiative) and which is not, nor why— even when such content contains irregularities or crimes, whether paid or unpaid. And where the company discloses its moderation criteria, we have observed serious inconsistencies in many cases where its own rules are not applied.

This opacity undermines public trust in the company’s genuine commitment to freedom of expression. After all, freedom is only effective when accompanied by transparency. Opacity in moderation allows freedom to be granted only to those chosen by the company, and even then it is done in a concealed manner.

The lack of adequate moderation, especially in advertising—which is how the company makes money—also severely compromises the safety of users and legitimate advertisers. An investigation published this month by the Financial Times revealed that Meta programs its algorithms to reduce the detection of irregularities and crimes in paid content. This is a hidden moderation criterion that causes significant material harm to users and has facilitated the spread of scams and fraud on Meta’s platforms. Public figures, institutions, private companies, members of parliament, and the Brazilian government itself have suffered from thousands of fraudulent ads that improperly use their images, brands, and reputations for illicit gain.

A recent report by Silverguard, a company specializing in digital financial protection, reveals an alarming figure: between January and June 2024, 80% of financial scams in Brazil originated on Meta’s platforms (WhatsApp, Facebook, and Instagram). The same report estimates that Brazilians lost 25 billion reais to financial scams in 2024—a figure that is underestimated, since half of scam victims do not file police reports. These numbers demonstrate the concrete impact of the company’s lack of interest and investment in moderation.

Moreover, the rhetoric of censorship frequently used by platforms conceals an important issue: the absence of transparency intended to exempt them from responsibility. Corporate discourse leads us to believe that censorship only comes by way of the State. However, in today’s reality, digital platforms constitute the main structure of user censorship on the Internet, unilaterally deciding what content will or will not be moderated. Even more concerning is the fact that these companies hold more information about their users than any nation-state has ever dreamed of holding about its citizens. They use people’s data—including sensitive data—to distribute personalized ads, regardless of whether they are legitimate, contain crimes of any kind, or place users at risk.

Without adequate transparency, freedom of expression becomes an empty promise.

On this point, we should clarify how Meta’s content moderation transparency reports work. Their structure is superficial and does not allow for independent auditing. We, as researchers, have repeatedly pointed out the lack of data for research purposes. Big tech companies have done what we call transparency-washing, releasing only the data that suits them. For example, companies disclose government and court requests for content removal. However, they remain opaque regarding their own day-to-day moderation practices, which account for 99.8% of content removed from platforms, according to the DSA Transparency Database.

It is important to highlight that Meta adopts a different behavior in Brazil compared to Europe, where it provides considerably more detailed moderation reports and offers data to researchers. This disparity reinforces the argument that the company is capable of offering greater transparency but opts to neglect Brazil.

The absence of regulation for digital platforms will worsen the already opaque scenario in which we find ourselves. Without transparency in content moderation, we are standing before a “private and secret tribunal” operated by the platforms themselves, which have the power to control the public debate according to their commercial interests, while using the rhetoric of freedom to maintain their immunity from local laws.

There is an urgent need for a regulatory framework and normative guidelines that can ensure transparency mechanisms for digital platforms in Brazil, in order to protect fundamental rights and the health of our democracy.