Platforms must now hunt for illegal content without waiting for a judge
Em um momento em que as plataformas digitais se tornaram praças públicas de alcance global, o Supremo Tribunal Federal do Brasil decidiu que a neutralidade passiva não é mais uma opção aceitável. A partir da publicação do acórdão, empresas como Meta, Google e X terão sessenta dias para remover proativamente conteúdos ilegais — terrorismo, racismo, incitação ao suicídio, ataques à democracia — sem aguardar ordem judicial. A decisão não é apenas técnica: é uma declaração de que o poder de moderar o discurso público não pode residir exclusivamente nos tribunais, mas também nas próprias mãos de quem lucra com ele.
- O STF virou a lógica anterior de cabeça para baixo: plataformas não podem mais esperar uma ordem judicial para agir — agora têm obrigação de buscar e remover conteúdos ilegais por conta própria.
- A indústria de tecnologia pediu seis meses para se adaptar; o tribunal concedeu sessenta dias, sinalizando que a urgência da proteção democrática supera a conveniência operacional das empresas.
- O risco de overcorrection paira sobre o debate: o ministro André Mendonça alertou que responsabilizar plataformas por conteúdo de usuários pode levá-las a remover discursos legítimos por excesso de cautela.
- Uma salvaguarda ainda em análise pode proteger plataformas que removerem conteúdo de boa-fé após análise séria — mas sua forma final ainda não está definida, deixando margem de incerteza.
- A decisão se apoia em precedente de junho passado e endurece obrigações já existentes, transformando compromissos difusos em prazo concreto com consequências reais para o descumprimento.
O Supremo Tribunal Federal impôs nesta semana uma virada significativa nas regras que governam as grandes plataformas digitais no Brasil. A partir da publicação formal do acórdão, prevista para a próxima semana, Meta, Google e X terão sessenta dias para remover ativamente conteúdos ilegais — incluindo material terrorista, discurso de ódio racial, incentivo ao suicídio e ataques às instituições democráticas — sem precisar aguardar determinação judicial.
A mudança é estrutural. Até então, o Marco Civil da Internet estabelecia que as plataformas só respondiam por conteúdo ilegal se descumprissem uma ordem de remoção emitida por um juiz. A nova decisão inverte essa lógica: a responsabilidade de identificar e agir passa a ser das próprias empresas. O relator Dias Toffoli rejeitou o pedido da indústria por seis meses de prazo, considerando sessenta dias suficientes. O presidente da corte, Edson Fachin, sugeriu uma salvaguarda: plataformas não deveriam ser punidas por remover conteúdo que, após análise séria, não puderam classificar com certeza — mas esse ponto ainda está em deliberação.
O debate interno no tribunal revelou tensões reais. Flávio Dino defendeu agilidade, argumentando que a demora gera insegurança jurídica num setor em constante transformação. Já André Mendonça, embora favorável ao prazo de sessenta dias, alertou para o risco oposto: plataformas excessivamente cautelosas podem acabar silenciando discursos que deveriam ser protegidos. A pergunta que fica é se duas semanas são tempo suficiente para que sistemas de moderação em escala global sejam calibrados com precisão suficiente para distinguir o ilegal do apenas incômodo.
Brazil's Supreme Court handed down a decision this week that will reshape how the country's largest tech platforms police their own digital spaces. On Thursday, the court set a sixty-day window for companies like Meta, Google, and X to begin actively removing content deemed illegal—material promoting terrorism, inciting racial hatred, encouraging suicide, or attacking democratic institutions. The clock starts when the court publishes its formal ruling, expected next week.
This is not a small adjustment. Until now, Brazilian law held that platforms bore responsibility for illegal posts only if they ignored a court order to take them down. The new ruling flips that logic. Companies must now hunt for and remove certain categories of harmful content on their own initiative, without waiting for a judge to tell them to do so. The shift reflects a growing global tension: platforms have become so vast and consequential that leaving moderation entirely to the courts has become untenable.
The tech industry had asked for at least six months to make the transition. Dias Toffoli, the justice who guided the case, dismissed that request. Sixty days, he said, is reasonable and more than sufficient. The court's president, Edson Fachin, suggested one concession: platforms should not be held liable if they remove content they genuinely could not determine was illegal, provided they conducted a serious analysis first. That safeguard is still being reviewed.
The decision builds on a ruling the court issued last June, which already expanded platform obligations under Brazil's Internet Civil Rights Framework. That earlier judgment widened the net of what counts as content platforms must address. Now the court is making those obligations stick with a hard deadline and real consequences for non-compliance.
Not everyone on the bench agreed on every point. Flávio Dino argued for moving quickly, warning that delay itself creates legal chaos in a fast-moving industry. André Mendonça supported the sixty-day timeline but raised a different concern: he worried that holding platforms jointly liable for user-generated content could backfire, making them so cautious that they remove legitimate speech. The risk, he suggested, is overcorrection.
What the court has essentially done is place the burden of content moderation squarely on the platforms themselves, at least for the most egregious categories of speech. No more hiding behind the excuse that they did not know, or that they were waiting for official instruction. The question now is whether sixty days is truly enough time for companies to build the systems, train the moderators, and establish the policies needed to catch and remove content at scale—and whether those systems will be precise enough to avoid silencing speech that should remain protected.
Citações Notáveis
60 days are a reasonable and more than sufficient timeframe for the necessary adjustments— Justice Dias Toffoli
The longer implementation is delayed in the technology world, the more outdated the ruling becomes— Justice Flávio Dino
A Conversa do Hearth Outra perspectiva sobre a história
Why did the court feel it needed to step in at all? Wasn't the old system—where platforms only had to respond to court orders—already working?
It was working in theory, but not in practice. By the time a court order came down, harmful content had already spread to thousands or millions of people. The court decided that waiting for the judicial system to move was too slow for the speed of digital harm.
So the court is essentially saying platforms should be the first line of defense, not the last.
Exactly. And that's a significant shift in responsibility. Platforms now have to develop their own judgment about what is illegal, not just follow orders from judges.
The tech companies asked for six months. Why did the court think sixty days was enough?
The justice who wrote the opinion said these are large, well-resourced companies with sophisticated systems already in place. He saw the six-month request as stalling. But there's a real question underneath: is sixty days actually enough to build compliant systems, or is the court being unrealistic about what implementation takes?
What happens if a platform removes something it shouldn't have—something that was actually legal speech?
That's the tension Mendonça raised. The court suggested platforms get some protection if they made a good-faith effort to analyze the content. But the details are still being worked out. That uncertainty is part of what makes this moment fragile.