The old shield is largely gone now.
Em um momento em que as fronteiras entre plataforma e editor se tornam cada vez mais tênues, o Supremo Tribunal Federal do Brasil impõe às grandes empresas de tecnologia um prazo de sessenta dias para reconfigurar suas práticas de moderação de conteúdo. A decisão, conduzida pelo ministro Dias Toffoli em junho de 2025, desfez décadas de proteção legal que exigia ordem judicial antes de qualquer responsabilização das plataformas. O que está em jogo não é apenas conformidade técnica, mas a pergunta mais profunda sobre quem, afinal, é o guardião do discurso democrático na era digital.
- O STF derrubou o escudo legal que protegia plataformas como Facebook e Google de responsabilidade por conteúdo de terceiros, exigindo adaptação imediata de sistemas de moderação.
- Doze recursos de empresas de tecnologia e aliados jurídicos circulam no tribunal, cada um tentando redefinir os limites do que a nova regra realmente exige.
- O Facebook pediu seis meses de prazo e responsabilidade restrita a conteúdos manifestamente criminosos — um sinal claro de que o setor considera o prazo de sessenta dias tecnicamente inviável.
- A proposta de Toffoli abre uma brecha sutil: plataformas podem remover certos conteúdos por demanda escrita, sem esperar ordem judicial, ganhando autonomia — e também responsabilidade — sobre seu próprio julgamento.
- Conteúdos amplificados artificialmente para manipular o debate público entram em zona cinzenta: plataformas escapam de punição automática se provarem que agiram com diligência e rapidez.
- O relógio de sessenta dias corre, mas a verdadeira prova virá quando as empresas tentarem implementar regras que ninguém ainda compreende por completo.
O Supremo Tribunal Federal do Brasil deu às maiores plataformas digitais do país sessenta dias para reformular suas práticas de moderação de conteúdo. O prazo foi proposto pelo ministro Dias Toffoli durante audiências em curso nesta semana e representa uma virada significativa nas expectativas do tribunal em relação a empresas como Facebook e Google — qualquer plataforma com mais de um milhão de usuários brasileiros está no escopo da decisão.
A pressão tem raiz em uma decisão histórica tomada em junho de 2025. Por anos, a legislação brasileira protegia as plataformas de responsabilidade enquanto removessem conteúdo ilegal após ordem judicial. O STF concluiu que esse modelo deixava posts prejudiciais circulando por tempo demais, corroendo o debate democrático e expondo direitos fundamentais. O antigo escudo, entenderam os ministros, não se sustenta num mundo em que essas plataformas moldam o que milhões de brasileiros veem e acreditam.
As empresas reagiram com uma série de recursos — doze ao todo — cada um sondando os limites práticos das novas regras. O Facebook foi o mais direto: pediu seis meses de prazo e responsabilidade restrita a conteúdos manifestamente criminosos. Google e outras companhias levantaram preocupações semelhantes sobre o ônus operacional do novo padrão.
A proposta de Toffoli busca um equilíbrio. Para crimes contra a honra — calúnia, injúria, difamação —, o sistema anterior se mantém em grande parte: uma ordem judicial ainda é necessária na maioria dos casos. Mas há uma abertura: plataformas podem remover certos posts mediante demanda escrita, sem aguardar decisão judicial. É uma mudança sutil que amplia o poder de julgamento autônomo das empresas.
No terreno mais complexo dos conteúdos amplificados artificialmente para manipular o debate público, Toffoli oferece uma saída: plataformas não enfrentam punição automática se demonstrarem que agiram com cuidado e rapidez ao identificar e remover posts ilegais. A prova do esforço fica a cargo da empresa. Um ponto permanece fixo: veículos jornalísticos estão isentos dessas novas regras, respondendo apenas à lei de direito de resposta — distinção que reconhece que plataformas e imprensa operam de formas fundamentalmente diferentes.
As audiências continuam. O verdadeiro teste virá quando as plataformas tentarem colocar em prática regras que, até agora, ninguém compreende por inteiro.
Brazil's Supreme Court has given the country's largest tech platforms sixty days to overhaul how they handle content moderation. The deadline, proposed by Justice Dias Toffoli during ongoing hearings this week, marks a sharp turn in how the court expects Facebook, Google, and other digital giants with more than a million Brazilian users to police what gets posted on their services.
The pressure comes from a seismic decision the court made last June. For years, Brazilian law had protected platforms from liability as long as they removed illegal content once a judge ordered them to. That shield is largely gone now. The court concluded that waiting for a judicial order left too many harmful posts circulating unchecked, corroding the democratic conversation and leaving people's fundamental rights exposed. The old rule, the justices reasoned, no longer fit a world where these platforms shape what millions of Brazilians see and believe.
Now the companies are fighting back—or at least asking for clarification. Twelve separate appeals from tech firms and their legal allies are working through the court system, each one probing the edges of what the new rules actually mean. Facebook has been the most explicit: give us six months, not two, the company argued. And only hold us responsible for content that is obviously, manifestly criminal. Google and others have raised similar concerns about the practical burden of the new standard.
Toffoli's proposal, unveiled as he cast his vote this week, tries to thread a needle. For crimes against a person's reputation—slander, libel, defamation—the old system mostly stays in place. A judge still needs to order removal in most cases. But there is a crack in the door: platforms can now take down certain posts based on a written demand alone, without waiting for a court. The shift is subtle but real. It gives platforms more power to act on their own judgment, at least in some situations.
The trickier terrain involves content that has been artificially amplified to manipulate public debate. Here, Toffoli's language offers platforms a lifeline. They will not face automatic punishment if they can show they acted carefully and moved quickly to remove or block posts they knew were illegal. The burden is on the company to prove it tried. This opens space for platforms to argue they did their best, even if harmful content spread for a time.
One boundary remains firm: news organizations and outlets whose primary business is journalism are exempt from these new rules. They answer only to Brazil's right-of-reply law, a separate framework the court has already blessed as constitutional. The distinction matters because it means the court is not treating Facebook the same way it treats a newspaper—a recognition that platforms and publishers operate differently, even when both distribute information.
The hearings continue. Toffoli was set to finish presenting his full reasoning on Thursday. What happens next will depend on how the other justices respond to his proposal and how the tech companies adapt—or resist. The sixty-day clock is ticking, but the real test will come when platforms try to implement rules that no one has fully figured out yet.
Citações Notáveis
The old rule left too many harmful posts circulating unchecked, corroding democratic conversation and leaving people's fundamental rights exposed.— Brazil's Supreme Court reasoning in June 2025 decision
Platforms can avoid punishment if they prove they acted carefully and moved quickly to remove or block illegal content.— Justice Dias Toffoli's proposal
A Conversa do Hearth Outra perspectiva sobre a história
Why did the court decide it needed to change the rules at all? What broke?
The old system required a judge to order removal before a platform could be held responsible. But by the time a judge acts, the damage is done—the post has spread, people have been harmed, the lie is already in circulation. The court decided that standard no longer protects democracy.
So now platforms have to guess what's illegal and take it down themselves?
Not quite. Toffoli's proposal keeps judges in the loop for most defamation cases. But it does give platforms room to act on their own for some content, and it lets them avoid punishment if they can show they tried hard and moved fast.
Facebook asked for six months instead of sixty days. Will they get it?
That's still being decided. The court is hearing arguments this week. But the tone suggests the justices are not inclined to give much more time. Sixty days is already a compromise.
What about content that is just false but not technically criminal?
That is the gray zone Toffoli did not fully resolve. The rules focus on illegal content—crimes, defamation, that sort of thing. Misinformation that is not criminal sits in a murky space.
Does this apply to news websites too?
No. Newspapers and news organizations are carved out. They follow different rules. The court is treating platforms differently from traditional media, which makes sense because they operate in fundamentally different ways.
What happens if a platform ignores the sixty-day deadline?
That is not spelled out yet. But the court has shown it is willing to fine and pressure these companies. Noncompliance would likely trigger sanctions.