A system of checks and balances for a digital world that didn't exist when the original law was written
The STF ruled in 2025 that Article 19 of Brazil's Internet Civil Rights Law is partially unconstitutional, expanding platform liability for user-generated illegal content beyond court-ordered removals. Toffoli's proposal requires platforms with over 1 million Brazilian users to implement 'duty of care' standards, transparency reporting, and dedicated support channels within 60 days of judgment publication.
- Supreme Court ruled in June 2025 that Article 19 of Brazil's Internet Civil Rights Law is partially unconstitutional
- Platforms with over 1 million Brazilian users have 60 days to implement new content moderation duties
- New rules require 'duty of care' for terrorism, child abuse material, suicide incitement, and related crimes
- Meta and Google filed objections claiming the ruling's language is too vague for consistent enforcement
Brazil's Supreme Court is reviewing appeals against expanded digital platform liability rules, with Justice Toffoli proposing a 60-day implementation deadline for major tech companies to comply with content moderation obligations.
Brazil's Supreme Court opened hearings this week on a case that could reshape how the country's largest tech platforms handle illegal content. The court is reviewing appeals against a decision it made last year that fundamentally changed the rules: platforms can now be held liable for user-posted material that violates the law, even without a court order telling them to take it down. Justice Dias Toffoli, who is steering the case, laid out his position on Thursday, proposing a 60-day window for the biggest companies to get into compliance.
The original ruling, handed down in June 2025, declared part of Brazil's Internet Civil Rights Law unconstitutional. That law had essentially shielded platforms from responsibility for what their users posted, as long as they removed content when a judge ordered them to. The court decided that shield was too broad. Now platforms face what amounts to a duty of care: they must actively work to prevent the spread of the most serious illegal material—terrorism, child sexual abuse material, incitement to suicide, and related crimes. They also have to publish annual transparency reports and set up dedicated channels where people can report problems, whether they use the platform or not.
Toffoli's 60-day deadline would begin counting from the moment the court publishes its final judgment. The rule applies only to large platforms, those with more than a million registered users in Brazil. Smaller services get a pass. The justice also proposed a practical exception: lawsuits filed before the court finishes its work can proceed under the old understanding, but everything after that point follows the new rules. This matters because it gives platforms a clear cutoff and protects people who sued before the change took effect.
One significant adjustment came from Facebook's pushback. The court had used the phrase "presumption of responsibility" when talking about paid ads and promoted content. Toffoli changed it to "presumption of relative fault," a shift he said would bring clarity and legal certainty. It's a small semantic move, but in law, words carry weight. The difference between absolute and relative responsibility could mean the difference between a platform being automatically liable or having room to argue its case.
Google and Meta are not finished fighting. Both companies filed objections claiming the court's language is vague enough to cause problems down the line. Meta wants the ruling to specify "manifestly" illegal content, arguing that without that word, platforms might remove material that isn't actually against the law, chilling legitimate speech. Google says the current wording could lead to wildly different interpretations in lower courts across the country, creating a patchwork of liability standards that makes compliance nearly impossible. They're asking for more objective, measurable criteria.
Toffoli pushed back on the censorship concern. He noted that anyone whose content gets removed can go to court and ask to have it restored, and the platform won't owe them damages for that process. He called it a system of checks and balances for a digital world that didn't exist when the original law was written. The court, he said, was being careful not to hand platforms a blank check to remove whatever they wanted, but also not to let them ignore serious harms.
The case sits at the intersection of two powerful forces: the platforms' need for legal certainty and the public's interest in preventing the spread of material that causes real damage. Brazil's courts will now watch how this plays out. If the platforms comply within 60 days, the question becomes whether their efforts actually reduce illegal content or simply shift the burden of moderation onto users and courts. If they don't comply, the court will have to decide what enforcement looks like. Either way, the next two months will show whether this ruling is a genuine shift in how Brazil regulates the digital sphere or another round in an endless negotiation between tech companies and government.
Citas Notables
This is not censorship. Anyone whose content is removed can go to court to have it restored, and the platform won't owe damages for that process. It's a model of checks and balances for this new world we're living in.— Justice Dias Toffoli
La Conversación del Hearth Otra perspectiva de la historia
Why did the court decide platforms needed to be liable for user content in the first place? Wasn't the old law designed to protect them?
The old law was written in 2014, before social media became the primary way Brazilians got information. It made sense then to shield platforms from liability as long as they obeyed court orders. But the court looked at what was actually happening—massive amounts of illegal content spreading before any judge could intervene—and decided the shield was too broad. They wanted platforms to take responsibility for the worst material without waiting for a lawsuit.
So why is 60 days enough time? These are global companies with millions of users.
That's exactly what the platforms are arguing. But Toffoli seems to be saying that the largest ones—the ones with a million users in Brazil—already have the infrastructure. They have trust and safety teams, they have automated systems. Sixty days is enough to point those systems at the specific harms the court named: terrorism, child abuse material, suicide incitement. It's not asking them to moderate everything, just the most serious stuff.
What happens if a platform removes something that isn't actually illegal?
That's Meta's worry, and it's legitimate. But Toffoli built in a safety valve. If you get your content removed and you think it was wrongful, you can sue. The platform won't automatically owe you money, but you get your day in court. It's not perfect—it puts the burden on the user to fight back—but it's not censorship in the sense that the platform has the final word.
Why does Google care about objective criteria?
Because Brazil has thousands of judges in different cities and states. If the Supreme Court's language is loose, each judge might interpret it differently. One court might say a post is manifestly illegal; another might say it's borderline. Google would face inconsistent liability across the country. They're asking for bright-line rules so they can actually comply.
Is this ruling likely to stick, or will the platforms keep fighting?
The court has already decided. What's happening now is the platforms are trying to shape how it gets enforced. They won't overturn the ruling, but they might get tweaks to the language or more time to comply. The real test comes in six months, when we see whether the platforms actually changed their behavior or just filed more lawsuits.