Pedophilia cannot exist on the network, regardless of your politics
Em um momento em que as instituições democráticas brasileiras ainda cicatrizam as feridas do 8 de janeiro, o Supremo Tribunal Federal sinaliza que a era da imunidade irrestrita das plataformas digitais pode estar chegando ao fim. O presidente da corte, Luís Roberto Barroso, revelou em Frankfurt que os ministros estudam duas exceções ao Marco Civil da Internet — a lei de 2014 que protegeu plataformas de responsabilidade por conteúdo de terceiros. A proposta não é apenas jurídica: é um reconhecimento de que a arquitetura da liberdade online, tal como foi concebida, não antecipou a escala do dano que ela poderia abrigar.
- O STF pressiona para responsabilizar plataformas por conteúdo criminoso e violações de direitos fundamentais mesmo sem ordem judicial prévia, rompendo com quase uma década de proteção legal.
- Os ataques golpistas de 8 de janeiro de 2023 e a proliferação de discurso de ódio nas redes sociais tornaram urgente o debate — o ministro Gilmar Mendes declarou o artigo 19 do Marco Civil 'obsoleto'.
- Dois casos concretos aguardam julgamento no STF: mulheres que tiveram perfis falsos ou comunidades ofensivas mantidos no ar por plataformas que se recusaram a agir, e qualquer decisão pode virar precedente vinculante.
- No Congresso, o PL 2630 está travado desde abril, e reformas eleitorais que aumentariam a responsabilidade das plataformas esbarram na regra de anterioridade eleitoral — deixando o caminho livre para o Judiciário agir.
- Barroso rejeitou a divisão ideológica em torno da regulação, argumentando que proteger crianças e a democracia não é pauta de esquerda nem de direita — é uma necessidade civilizatória.
O Supremo Tribunal Federal brasileiro está sinalizando uma virada histórica na regulação da internet no país. Em conferência realizada em Frankfurt, o presidente da corte, Luís Roberto Barroso, revelou que os ministros estudam duas exceções ao Marco Civil da Internet — a lei de 2014 que protege plataformas de responsabilidade por conteúdo gerado por usuários enquanto cumprirem ordens judiciais de remoção.
A primeira exceção obrigaria as plataformas a usar algoritmos para identificar e remover conteúdo criminoso de forma proativa, sem esperar determinação judicial. A segunda reduziria o limiar para violações de direitos fundamentais — como imagens íntimas não consensuais ou violações de direitos autorais — permitindo responsabilização a partir de uma notificação privada, sem necessidade de ordem judicial.
Barroso rejeitou a ideia de que regulação seja uma bandeira ideológica. Para ele, combater pedofilia e proteger a democracia transcende divisões políticas. Seu colega Gilmar Mendes foi mais direto: os ataques golpistas de 8 de janeiro de 2023 e a disseminação de desinformação e ódio nas redes expuseram a 'obsolescência' do modelo atual de responsabilização das plataformas.
Dois casos concretos já estão na fila do STF e podem estabelecer precedente vinculante. Em um deles, uma mulher teve um perfil falso criado em seu nome no Facebook, que se recusou a remover. No outro, uma professora pediu ao Orkut que derrubasse uma comunidade dedicada a atacá-la — sem sucesso. Ambas buscam remoção de conteúdo e indenização por danos morais. O ministro Dias Toffoli afirmou que um dos casos 'está pronto para julgamento, só aguarda data'.
Enquanto isso, o caminho legislativo permanece bloqueado. O PL 2630, que criaria um marco regulatório e um órgão fiscalizador, está parado na Câmara desde abril. Reformas eleitorais que aumentariam a responsabilidade das plataformas por conteúdo golpista esbarram na regra de anterioridade e não entrarão em vigor antes das eleições municipais do próximo ano. O sinal do STF é claro: se o Congresso não age, o Judiciário avança.
Brazil's Supreme Court is quietly reshaping how the internet works in the country, and the shift could upend nearly a decade of legal protection for social media platforms. Speaking at a conference in Frankfurt this week, the court's president, Luís Roberto Barroso, revealed that the justices are weighing two significant exceptions to the Marco Civil da Internet—the 2014 law that has shielded platforms from liability for user-generated content as long as they comply with court removal orders.
The first exception would require platforms to deploy algorithms capable of identifying and removing criminal content without waiting for a judge's instruction. Barroso was blunt about the necessity: with the sheer volume of posts flowing across social networks every second, human moderation is simply impossible. The algorithm itself becomes the gatekeeper. The second exception would lower the bar for violations of fundamental rights—cases involving non-consensual intimate images, copyright infringement, or similar harms—allowing platforms to be held liable based on a private notification alone, without a court order.
Barroso framed this not as judicial overreach but as civilizational necessity. He pushed back against what he called an artificial ideological divide, where conservatives supposedly oppose regulation and progressives support it. "Pedophilia cannot exist on the network," he said, "regardless of whether you are liberal, conservative, or progressive." The court president acknowledged the complexity of calibrating such rules, but he left no doubt about the direction: regulation is now "imperative and indispensable."
His colleague Gilmar Mendes went further, directly attacking the current law's relevance. The proliferation of hate speech, disinformation, and attacks on democratic institutions through social media—culminating in the January 8, 2023 coup attempt—has exposed what Mendes called the "obsolescence" of Article 19 of the Marco Civil. The court, he argued, needs to rethink the entire regulatory model to hold platforms properly accountable.
Two cases are already queued before the Supreme Court, either of which could establish binding precedent. In one, a woman created a fake profile impersonating her and hurling insults at others; Facebook refused to take it down. In the other, a teacher asked Orkut (now owned by Google) to remove a community dedicated to mocking and attacking her; the platform declined. Both women are seeking removal of the content and damages for emotional harm. These cases were initially scheduled for June but were postponed. Justice Dias Toffoli, who is handling one of them, said this week that "the case is ready for judgment, just waiting for a date."
Barroso was careful to note that the court has not yet issued an official ruling, and that the justices are still deliberating. But the signal is unmistakable. Meanwhile, legislative efforts to tighten platform accountability have stalled. Bill 2630, which would create new rules and establish a regulatory body to oversee them, has been stuck in the Chamber of Deputies since April, deadlocked over which agency should supervise enforcement. A new electoral code and a mini-electoral reform that would also increase platform responsibility for coup-related content have run into the one-year rule—they cannot take effect before next year's municipal elections, making them effectively dead for now. The court's move suggests that if Congress cannot act, the judiciary will.
Citações Notáveis
Pedophilia cannot exist on the network, regardless of whether you are liberal, conservative, or progressive.— Luís Roberto Barroso, STF president
The obsolescence of Article 19 of the Marco Civil is evident, showing the need to rethink the regulatory model to properly hold platforms accountable.— Gilmar Mendes, STF justice
A Conversa do Hearth Outra perspectiva sobre a história
Why does the court think algorithms can do what humans cannot?
The volume is the answer. There are millions of posts per minute across all platforms combined. No human team could review them all. An algorithm, if properly trained, can flag criminal content instantly—child exploitation, threats of violence, that sort of thing. The court is saying: if you can build an algorithm to recommend content that keeps people scrolling, you can build one to catch crimes.
But who decides what the algorithm removes? Doesn't that give the platform enormous power?
That's the tension no one is quite resolving yet. The court is saying the platform has to do it, but it's not saying the platform gets to decide what counts as criminal. That's still defined by law. The platform just has to enforce it automatically instead of waiting for a judge to tell them.
What about the second exception—fundamental rights violations?
That's broader and vaguer. It includes things like revenge porn, which Brazil already handles through private notification. But "fundamental rights" could stretch to cover harassment, defamation, privacy violations. The court is saying: don't wait for a lawsuit. If someone notifies you privately that their rights are being violated, take it down.
Why did the court feel compelled to act now?
January 8th. The coup attempt. Hate speech, disinformation, calls for violence—all of it spread across social media. The court looked at that and decided the 2014 law was written for a different internet. Back then, platforms were smaller, the stakes felt lower. Now they're infrastructure. They shape what people believe.
What happens to the platforms?
If the court rules this way, they become liable in ways they haven't been before. They can't hide behind the excuse that they didn't know about the content. They have to actively police it. That's expensive. It's also risky—algorithms make mistakes, and if a platform removes something it shouldn't have, it could face its own lawsuits.
Is Congress just giving up?
Not intentionally. But yes, effectively. The big regulatory bill has been stuck for months. Electoral reform can't happen in time for next year's elections. The court is stepping in because the legislature can't move.