Judicial activism or democratic safeguard? The new limits of digital electoral propaganda in Brazil
General context
On March 1, 2024, the Superior Electoral Court of Brazil (hereinafter TSE) published 12 resolutions that will govern the 2024 elections to elect mayors and councilors of the 5.568 municipalities in the country for the period 2025-2028.
Brazilian electoral resolutions are regulations issued in each electoral process and aim to regulate federal legislation on elections (Article 105 of the Electoral Law), adapting general rules to specific contexts. They refer to various topics, such as electoral surveys and candidacy records. For example, they include the calendar that will govern the year's activities and the updated values for the distribution of the Special Campaign Financing Fund. Since they come from an organ of the judiciary, it is understood that these devices should not introduce legislative innovations. This text addresses some of the most relevant points of Resolution No. 23.732/2024, on electoral propaganda, which seeks to adapt the rules that govern campaign events to the new realities of the digital environment.
A note on the process
Before delving into the merits of the norm itself, it is important to make a brief comment on the formalities for drafting and approving TSE resolutions. This process is carried out by the office of the vice president of the Court and is always preceded by publishing the drafts of the texts. These first versions are subjected to public hearings, where the TSE receives written and oral contributions from various sectors of society. After these contributions, the vice presidency's office drafts the final texts and takes them to a vote in the plenary. When the regulations are passed, they are then published and are valid for the next electoral process.
The particularities of the 2024 process
According to tradition, in 2024, the TSE published draft texts of the resolutions and organized public hearings to receive comments and suggestions from interested parties from all sectors. According to the Court itself, the sessions reached a record participation. Between civil society groups, public sector organizations, and digital platform teams, the process entailed 80 oral and 945 written contributions.
However, when the final text was published, the organizations that participated in the process were surprised with a provision that was not in the previous versions. Along with other changes, Article 9-E implemented a controversial modification to the liability regime for intermediaries during the electoral process that will be described below.
Relevant points of the new text on electoral propaganda in the digital sphere
Change in the liability framework of intermediary platforms – Article 9º-E
As mentioned, the most controversial clause of the resolution is art. 9º-E that modified the intermediary liability regime established by the Marco Civil de Internet (MCI) [Internet Civil Framework]. The wording of article 19 of the MCI establishes that platforms will only be civilly liable for content published by third parties when they do not comply with a judicial notice of removal, to avoid improper monitoring and removals by providers. Furthermore, the Electoral Law, in its article 57-F, repeats the terms of the MCI, reproducing its model of liability for electoral propaganda.
However, article 9-E introduces the possibility of civil and administrative liability for these companies when they do not immediately eliminate five groups of content considered as "risk cases." These are: (i) undemocratic behaviors, (ii) dissemination of disinformation, (iii) threats to the integrity of institutions and their servers, (iv) hate speech, (v) failure to comply with the obligation to label artificial intelligence content, wich was introduced by this resolution.
This new clause alarmed various actors involved in the discussion mainly for two reasons:
- The first: it is a regulatory instruction issued by a Court that expressly contradicts federal laws, in this case, the Electoral Law and the Internet Civil Framework. The purpose of the resolutions of the Electoral Judicial Branch is to regulate already existing provisions, not to modify parameters established by the legislative branch. Two intermediary liability regimes cannot coexist in the Brazilian legal system.
- Furthermore, in terms of its practical application, the consequence of the resolution is precisely what the MCI tries to avoid: private censorship by the platforms. Once they incur the risk of civil and administrative liability, companies may choose to secure and preserve their own services and operations, to the detriment of their users' freedom of expression, as well as the free circulation of speech. This is a perverse consequence of the resolution because it encourages providers to act as curators of public debate, particularly during the electoral period when the dissemination of content, whether it be news, political propaganda, opinions, or others, reaches its maximum democratic value.
Other provisions of the resolution
In addition to the change in the liability regime, Resolution No. 23.732/2024 included new provisions with obligations for candidates, political parties and platforms. Article 9º-C, for example, incorporates into the regulations the express prohibition of disseminating disinformation content, with the following wording: "facts that are notoriously false or decontextualized with the potential to cause harm to the balance of the election or the integrity of the electoral process." Article 9º-C §1º, for its part, specifically prohibits the use of deep fakes, a concept defined as "synthetic content in audio, video or combination thereof, that has been digitally generated or manipulated, even with authorization, to create, replace or alter the image or voice of a living, deceased, or fictional person."
Article 9-D expressly establishes the prohibition of sponsoring content that contains disinformation on social media platforms. This device is relevant because Brazilian legislation authorizes paid political propaganda on digital platforms (article 57-C of the Electoral Law) through the “content boost” mechanism. Additionally, the same article prescribes various obligations for platforms, such as the development of terms of use aimed at reducing the circulation of disinformation, the implementation of reporting channels, the publication of transparency reports disclosing actions taken to improve content recommendation systems, the conduct of impact assessments of services on electoral integrity, among others.
The resolution also introduces changes related to the use of artificial intelligence. Article 9-B regulates its use in electoral campaigns by ordering explicit labeling when the tool is used. For example, it establishes that audio and video pieces must be accompanied by a prior notice regarding the use of the technology, and requires the use of watermarks and audio description in images and videos. The resolution also establishes an exception to the application of such labels, in cases of adjustments intended to improve the quality of the image or sound, the production of visual identity and brands, and for propaganda resources commonly used in campaigns, such as photo montages.
Lastly, the regulation represents a step forward in transparency by inaugurating a public library of advertisements, an updated repository that compiles, in real time, the propaganda pieces, the validity period of the advertisement, the values, and the characteristics of the users reached by the contracted advertising. So far, apart from campaign expense statements, in which candidates report on their expenses during the electoral period, no instrument revealed information about the reach and audience of these sponsored posts.
Conclusion
The publication of Resolution No. 23.732/2024 sparked extensive debates among the various actors involved in the discussion on platform regulation and its intersection with elections in Brazil. On the one hand, it addresses long-standing societal demands by introducing transparency obligations for platforms. This is evident in the advertisement repositories, where the electorate can track spending on propaganda on platforms by candidates and parties. Additionally, represents an initial regulatory approach to artificial intelligence, with reasonable provisions on, for example, the need to label content that uses this tool.
On the other hand, the regulations introduce a rule into the system that has the potential to violate the right to freedom of expression on the Internet, generating more doubts than confidence and legal certainty. Furthermore, it crosses dangerous lines regarding the competence of a judicial body to introduce legislative innovations that go against federal precepts. Although the challenge of combating disinformation is giant, it cannot lead to the flexibility of federal processes and laws.
It is also important to remember that since 2020, Brazil has been immersed in an ongoing discussion process of a bill (PL 2630 / 2020) aimed at updating the regulatory framework of the digital environment, introducing various transparency obligations and systemic actions to promote safe, plural, and diverse spaces for expression. However respectable the intentions of the TSE may be and however enduring and complex the legislative process, this branch of government has the power and authority to issue regulations of this nature; where a solid, legitimate, and effective legal framework must emerge to address the challenges presented by the digital ecosystem.
