Roundtable Summary | Intermediary Liability in Brazil

Summary of the Comparative Legislation Roundtable on Intermediary Liability in Brazil.

1. Background

The Marco Civil da Internet (Law No. 12,965/2014) established Article 19 as the baseline for intermediary liability in Brazil, generally shielding platforms from liability for third-party content unless they failed to comply with a judicial takedown order. After roughly a decade without congressional reform, the Supreme Federal Court (STF) ruled in 2025 (Case No. 987) that Article 19, on its own, provided insufficient protection for fundamental rights and was unconstitutional by omission. The Court therefore established a temporary four-regime framework pending legislative action. In clarifying rulings issued in June 2026, the STF extended the Article 19 exemption for “low-interference” providers beyond email services and confirmed that the Executive Branch could regulate the implementation and enforcement of the framework by decree. 

2. The four liability regimes

Regime 1 — Court-order safe harbor — Article 19 remains in effect for selected cases: A specific court order remains the trigger for liability in cases of defamation or offenses against honor, as well as for protected interpersonal communications and other services that do not interfere with the flow of information. 

Regime 2 — Notice & action - General crime or unlawful act: After receiving a valid notice, providers must assess the situation and take action. Liability follows the logic of Article 21, with the “reasonable doubt” standard preserved following a qualified analysis. This is the general regime for individual content. 

Regime 3 — Due Diligence I - Systemic Failure / Serious Crimes: For the exhaustive list of serious crimes, liability hinges on systemic failure: whether adequate measures were taken to prevent or remove the relevant content. Isolated instances of content are not sufficient. 

Regime 4 — Due Diligence II - Paid Ads & Artificial Distribution: A relative presumption of fault applies to illegal content disseminated through paid ads, promotion, or artificial distribution networks. The provider can rebut this presumption by proving diligent and timely action. 

3. What the Decrees Actually Do

They transform the STF’s legal principle from a liability ruling into an administrative and operational framework. 

I. Translating the thesis into procedures

Decree 12.975/2026 specifies notice requirements, acknowledgment, reasoned decisions, appeals, safeguards against abusive notices, and the “reasonable doubt” standard. 

II. Separate ex post liability from ex ante governance

It gives concrete form to systemic-failure duties: risk monitoring, prevention, transparency, record-keeping, and evidence that adequate safeguards were actually implemented. 

III. Differentiate monetized distribution

Articles 16-K–16-M address paid advertising and boosting: preventive contracting measures, a rebuttable presumption of liability, and the retention of ad and advertiser information. 

IV. Create a regulatory center of gravity

Art. 19-A assigns the regulation, supervision, and investigation of provider duties to the ANPD; Decree 12.976/2026 applies the same logic to violence against women in the digital environment. 

Bottom line: the Decrees serve as the bridge between the civil liability standards established by the STF and the compliance and enforcement mechanisms to be developed by ANPD. 

4. ANPD’s next move: defining the compliance standard

The public consultation process is where the open framework of the Decrees becomes operational law. 

I. Define the core concepts

“Duty of care/diligence,” “systemic failure,” “systemic risk,” “reasonable doubt,” and “identical content.” 

II. Establish evidence and metrics

What documentation, baselines, samples, indicators, and qualitative context can demonstrate adequate systems—or systemic failure? 

III. Calibrate by risk and service

Apply proportionality based on provider size, level of interference, state-of-the-art technology, and service-specific risk; preserve distinctions between regulatory regimes.

IV. Design enforcement & information flows 

Standardize reporting, confidentiality, notices and appeals, requests from regulators, and coordination with other authorities; prioritize responsive and graduated oversight.

5. Two Open Questions Raised During the Discussion

The discussion highlighted two unresolved questions regarding the new framework. First, it was questioned whether the regime is ultimately means- or results-oriented: does compliance depend primarily on whether providers adopt reasonable procedures, or will liability hinge on the volume and outcomes of content moderation? In particular, the concept of “systemic failure” implicitly requires defining what would constitute “systemic success.” 

Second, concerns were raised about scope creep. Ancillary obligations relating to researcher access to data, transparency reporting, and risk-assessment disclosures could potentially push the framework beyond traditional intermediary liability and toward a broader form of platform regulation.

6. Audience Discussion 

The audience discussion focused on several unresolved questions surrounding the STF’s new intermediary liability framework. Participants explored whether Regime 3 resembles the EU Digital Services Act’s risk-based approach, particularly regarding the concept of “systemic risk,” and questioned whether risk assessments are limited to the framework’s closed list of crimes. The discussion also examined the relationship between the new paid-advertising regime and Brazil’s existing consumer-liability rules for marketplaces, including when a marketplace may be held responsible for promoted listings. 

Participants also raised concerns about the distinction between “non-interference” and active content moderation, questioning whether treating non-interference as inherently safer could create perverse incentives—for example, making highly uncurated platforms less exposed than platforms that actively moderate content. This prompted a discussion of the risk of over-removal, the STF’s explicit rejection of strict liability, and whether safeguards such as the “reasonable doubt” defense will be effective in practice. The evolution of the non-interference doctrine and the ANPD’s authority to tailor obligations according to the type and size of the service were also discussed. 

The discussion further addressed enforcement and institutional authority. Participants questioned the Executive Branch’s role in issuing decrees to implement a framework established by the STF, whether such an approach could be challenged, and how “systemic failure” would be established in practice. The still-uncertain division of responsibilities between the ANPD—which is expected to develop technical standards—and the courts—which retain authority over liability determinations—emerged as a key issue. The treatment of private messaging and public-facing features was also discussed, with private communications remaining under the Article 19/16-P framework while features such as channels may fall under the platform-like regime. 

Finally, participants discussed the political and institutional future of the framework, particularly in light of Brazil’s upcoming elections and the possibility of changes to the Executive’s decrees. The discussion highlighted the tension between concerns over decree-based regulation and industry demands for greater regulatory certainty, while noting that challenges to the decrees are already pending before the STF.

7. Disclaimers

Chatham House Rule: This report adheres to the Chatham House Rule throughout.

AI disclaimer: Artificial intelligence tools were used to help prepare this summary. The text was reviewed prior to circulation.

Summary drafted by Josefina Gálvez, Research Assistant, CELE.