Brazil’s Supreme Federal Court ruled 8–3 that social-media platforms are civilly liable for illegal content uploaded by users if they fail to remove it after receiving an extrajudicial notice from victims or the public. The decision partially invalidates Article 19 of the 2014 Marco Civil da Internet, which had shielded providers unless they ignored a court order. Under the new standard, companies such as Meta, Google, X and TikTok must act promptly against posts that promote terrorism, anti-democratic actions, racism, misogyny, offences against children or other serious crimes, or face damages claims. For defamation and other crimes against honour, a judicial order remains necessary, but repeated reposts of content already judged unlawful may be taken down on notice alone. The justices introduced a “systemic failure” test that can trigger liability when platforms do not adopt preventive moderation or transparency measures. Providers must keep a legal representative in Brazil, publish annual transparency reports and maintain user complaint channels. Because the ruling carries binding effect, every court in Brazil must follow it until Congress enacts comprehensive legislation. Google and industry groups said the move risks curbing free expression and digital-economy growth, while the government’s solicitor-general called it a landmark step to curb online abuse and protect democracy.