JupiteX Get the app
Politics & Governance21 Nov 2026 · about 7 min

Lessons from Brazil and Peru on how to build AI-ready institutions across Latin America

The brief

AI-ready institutions can turn rules and principles into real oversight. The article frames readiness around three questions: Is authority clear? Can institutions handle new responsibilities? Can they spot and correct failures? These questions matter because a law alone does not ensure that anyone can implement it or hold decision-makers accountable. Brazil illustrates how capacity can grow before a general AI framework is settled. A proposal would make the National Data Protection Authority the default AI regulator, working alongside sectoral regulators. In February 2026, a separate law gave the authority full regulatory-agency status and greater technical, administrative, and financial authority. The article also points to transparency, review, participation, and accountability as developing safeguards. Peru has passed a general AI law and paired it with implementation mechanisms, while Brazil’s broader framework remained under legislative consideration in the text. The cases show different routes toward readiness, not a single finished model. For other countries, the lesson is to clarify responsibilities, build practical expertise and resources, and create channels to hear concerns and correct harm as AI is deployed.

01

What does it mean for a country to have AI-ready institutions?

AI-ready institutions can turn rules and principles into real oversight. The article frames readiness around three questions: Is authority clear? Can institutions handle new responsibilities? Can they spot and correct failures? These questions matter because a law alone does not ensure that anyone can implement it or hold decision-makers accountable.

Brazil illustrates how capacity can grow before a general AI framework is settled. A proposal would make the National Data Protection Authority the default AI regulator, working alongside sectoral regulators. In February 2026, a separate law gave the authority full regulatory-agency status and greater technical, administrative, and financial authority. The article also points to transparency, review, participation, and accountability as developing safeguards.

Peru has passed a general AI law and paired it with implementation mechanisms, while Brazil’s broader framework remained under legislative consideration in the text. The cases show different routes toward readiness, not a single finished model. For other countries, the lesson is to clarify responsibilities, build practical expertise and resources, and create channels to hear concerns and correct harm as AI is deployed.

02

How do Brazil and Peru differ in their approaches to AI regulation?

Brazil and Peru differ in how far their legal frameworks have advanced. Brazil is building institutional capacity while its general AI framework remains before a special committee in the Chamber of Deputies, with no vote date confirmed in the text. Peru, by contrast, is described as having passed a general AI law and established implementation mechanisms.

Brazil’s Senate approved an AI bill on December 10, 2024. A proposal attached to it would make the National Data Protection Authority the default AI regulator, alongside sectoral regulators. Peru’s 2023 law establishes principles that include data privacy, developing national AI talent, and cooperation on AI development at home. The provided text does not identify Peru’s implementation bodies or explain those mechanisms in detail.

The contrast is between institutional strengthening amid an unfinished broad framework in Brazil and a law-plus-implementation approach in Peru. Brazil also has a separate law that turned its data protection authority into a full regulatory agency in 2026. These examples suggest that countries can build capacity before or alongside a comprehensive law, while keeping authority and implementation arrangements clear.

03

Which institutions are taking on AI oversight in Brazil and Peru?

Brazil’s oversight is being taken up by several institutions rather than one settled, all-purpose regulator. The National Data Protection Authority, or ANPD, is being strengthened and could become the default AI regulator under a proposal, alongside sectoral regulators. Brazil’s Superior Electoral Court has also issued rules addressing AI in elections.

Ahead of the 2026 elections, the court issued fourteen resolutions, including restrictions on AI systems recommending candidates or signaling voting preferences. It also required disclosure of AI-generated or digitally manipulated campaign content and could enforce its rules by suspending profiles. Separately, a February 2026 law made ANPD a full regulatory agency with greater technical, administrative, and financial authority.

Peru is described as having passed a 2023 general AI law with implementation mechanisms, but the supplied text does not name the institutions responsible for those mechanisms. It therefore supports a clear account of Brazil’s institutional actors, but not a comparable list for Peru. That gap matters: identifying who leads, and how responsibilities are divided, is one of the article’s tests of effective governance.

04

How far have Brazil and Peru moved from AI plans to rules that can be implemented?

The two countries have moved beyond plans by different amounts. Peru is described as one of the first Latin American countries to pass a general AI law coupled with implementation mechanisms. The text says its 2023 law establishes principles including data privacy, national AI talent development, and cooperation on AI development at home.

Brazil’s general AI framework had not yet completed its legislative process: it remained before a special committee in the Chamber of Deputies, and no vote date had been confirmed. Still, the Senate approved a bill on December 10, 2024, and a separate law in February 2026 gave ANPD greater regulatory authority. The Superior Electoral Court also issued fourteen resolutions governing the 2026 electoral process, including AI restrictions.

So Peru has the more advanced general-law framework in the account, while Brazil shows that implementation rules and institutional capacity can develop before the broader framework is settled. The excerpt gives no details about Peru’s mechanisms, so it cannot establish how fully they are staffed or operating. Both cases leave the practical test of capacity and accountability central.

05

What can countries do to address specific AI risks before they have a dedicated AI regulator?

Countries do not have to wait for a dedicated AI regulator before addressing every risk. The article describes courts as possible interim regulators, especially when a risk falls within an existing institution’s area of responsibility. It also points to regional guidance and sectoral regulators as ways to clarify roles while broader laws are still being developed.

Brazil’s Superior Electoral Court offers a concrete example. Ahead of the 2026 elections, it issued fourteen resolutions covering the electoral process. These barred AI systems from recommending candidates or signaling voting preferences, and required disclosure of AI-generated or digitally manipulated campaign content. The court could enforce the rules by suspending profiles. Separately, the Inter-American Framework on Data Governance and Artificial Intelligence offers guidance countries can adapt.

The article suggests that countries without designated AI regulators could empower existing judicial institutions to address specific risks during a transition to an independent regulator. This is not presented as a replacement for broader oversight. Institutions still need clear responsibilities, the ability to enforce rules, and channels for participation and accountability. Local conditions and the expertise of cross-sector specialists also matter when shaping responses.

06

What harms can occur when AI oversight fails to account for local inequalities?

When oversight ignores local inequalities, AI systems can reinforce unequal treatment rather than address it. The article highlights this risk through Brazil’s public security and surveillance debates. It says regulation must account for local conditions, including structural racism, and provide ways for civil society to participate and hold institutions accountable.

São Paulo’s Smart Sampa city surveillance pilot was scrutinized by civil society from its presentation in 2022. It was found to disproportionately profile Black individuals. Brazil’s first AI proposal was also criticized for insufficiently including Indigenous voices in legislative debate and for failing to consider structural racism in public security. These examples show how both deployed systems and policy-making can overlook affected communities.

Brazil’s legislative debates later incorporated concerns about racial discrimination and other contextual issues, according to the text. The article argues that institutions need clear channels for participation, oversight, and accountability, and the flexibility to listen to specialists across sectors. As other countries in Latin America pilot AI for security purposes, locally grounded scrutiny can help institutions identify and respond to unequal effects.

07

Why does AI governance need to assign clear roles across the data lifecycle, from data collection to deployment?

AI governance needs clear roles because responsibilities stretch across the data lifecycle, from collection through later use and deployment. If no institution or actor knows what it is accountable for, oversight can become fragmented. The article treats clear authority as a basic test of whether institutions can implement rules and address failures.

The Organization of American States launched the Inter-American Framework on Data Governance and Artificial Intelligence, developed by multisector stakeholders and governments. One recommendation is to clearly define the processes, roles, and responsibilities of actors involved across the data lifecycle. Brazil has taken a step in this direction by assigning the National Data Protection Authority a proposed role as default AI regulator, alongside sectoral regulators.

The framework offers guidance countries can adapt even while drafting AI laws. That matters because institutional building need not wait for a comprehensive legal framework. Clear divisions can help regulators coordinate, understand where expertise is needed, and make accountability more visible. The article’s wider questions remain: whether institutions have enough capacity, and whether transparency, review, participation, and correction mechanisms develop alongside AI deployment.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web