TSE: Resolution goes against the Civil Framework – 03/10/2024 – Power

Jurists, representatives of big techs and civil society entities consulted by the Sheet believe that the resolution on electoral propaganda published by the Superior Electoral Court (TSE) on March 1st is illegal, as it would violate the Marco Civil da Internet.

The sticking point is article 9E. It establishes that internet platforms will be jointly and severally responsible “civilly and administratively when they do not promote the immediate unavailability of content and accounts during the electoral period”.

“Undemocratic” posts that violate certain legislation, including the Law on the Democratic Rule of Law, must be immediately removed; “notoriously untrue or seriously decontextualized facts” about the electoral process, “serious threat, direct and immediate, of violence or incitement to violence” against members of the Judiciary; “hateful behavior or speech,” including “racism, homophobia, Nazi, fascist, or hateful ideologies”; and “content manufactured or manipulated” by artificial intelligence without receiving the appropriate labels as required by the resolution.

The language of the article implies that companies may be held responsible for content that has not been reported by users or by the TSE and without there being a judicial order to remove the illicit post.

According to lawyers, this article changes the liability regime for big techs in Brazil, as anyone who finds infringing content on the platforms can sue the company, in addition to the author of the post.

The Marco Civil da Internet, the main law that regulates the sector in Brazil, from 2014, establishes that companies can only be civilly punished for third-party content if they do not remove it after a court order, except in cases of non-consensual nudity or property violation intellectual.

“This article cannot exist in a world where there is the Marco Civil da Internet – there is no such electoral exceptionalism, which goes against federal law”, says Carlos Affonso Souza, professor at the State University of Rio de Janeiro (Uerj) and director from the Institute of Technology and Society.

The resolution would also be violating the Brazilian Elections Law. This legislation establishes that providers can only be fined after judicial notification and that companies will only be responsible if they are proven to have “prior knowledge” of the publication.

According to the companies, if the new TSE resolution allows platforms to be punished for unreported content or without judicial notification, they will have to carry out a surveillance system, active monitoring of all publications during the electoral period.

As there are hundreds of millions of posts, companies would use, for the first filtering, artificial intelligence mechanisms to detect infringing content. But these mechanisms are still notoriously flawed —many use keyword searches— and there would inevitably be excessive removal, functioning as censorship.

“In the election, competing candidates will keep notifying non-stop, saying that it is content with hate speech, and the platforms will remove it”, says Souza.

Article 9E of the resolution has excerpts identical to the document with suggestions from minister Alexandre de Moraes delivered to the president of the Senate, Rodrigo Pacheco (PSD-MG) and the president of the Chamber, Arthur Lira (PP-AL) in April 2023, to be included in PL 2630, the fake news bill. Most of the suggestions from Moraes, who is now president of the TSE, were not included – and voting on the PL was suspended in May, due to lack of support.

According to Bia Barbosa, director of DiraCom (Right to Communication and Democracy and member of the Coalition Rights on the Network, the TSE “advanced the signal” after the omission of Congress, which was unable to approve the regulation proposal.

“We are all going to pay for the fact that Brazil did not regulate the liability regime”, he says.

“Yes, the Marco Civil is ten years old and we can review it – but the resolution is a risk to freedom of expression by determining the joint liability of platforms, which will remove content en masse.”

Lawyers and companies hope that the TSE will publish a clarification on the resolution, to specify whether they can be held responsible for content before judicial notification.

The companies believe that, in practice, the resolution equates internet platforms with media outlets such as newspapers and TVs in relation to responsibility for the content conveyed. In the same way that a newspaper can be sued for an article that was published, the platform could be sued for content. Moraes has already stated that “it must be equal in responsibility to communication and advertising companies, even with their peculiarities”.

The companies claim that they are mere distributors, and not publishers or content producers, so they cannot be held responsible.

In the resolution issued ten days before the second round of the 2022 election, called “police power”, Moraes has already expanded the TSE’s power to order platforms to remove certain content and impose fines if they do not comply with the orders within the specified period. According to Barbosa, if before it was the TSE deciding what content to remove, now, with the new resolution, the platforms will decide – and will choose to take down en masse to avoid punishment.

Legal experts highlight positive points of the resolution, such as the ban on deepfakes and chatbots that simulate conversations with candidates, the labeling of electoral content that uses artificial intelligence, and the institution of the so-called “duty of care”.

In article 9D, it is provided that platforms, in an election year, must prepare an assessment “of the impact of their services on the integrity of the electoral process”, in order to implement effective and proportionate measures to mitigate the identified risks.

It is something similar to the so-called “duty of care” provided for in the Digital Services Law that came into force in the European Union last August.

The resolution also includes a “Francichini clause” in the age of artificial intelligence. In 2021, the TSE revoked the state deputy from Paraná Fernando Francischini (at the time in the PSL) due to the publication of a video on the day of the 2018 elections in which he stated that the electronic voting machines had been rigged to prevent voting for the then presidential candidate Jair Bolsonaro.

The clause prohibits the use, in electoral propaganda, of content manufactured to disseminate untrue or out-of-contextual facts with the potential to harm the electoral process. Non-compliance constitutes abuse of political power and misuse of social media, which leads to revocation of registration or mandate.