AI Regulation Tracker / Legislation pending

Brazil's AI Act clears the Senate but stalls in the Chamber, final vote now expected in 2026

Projeto de Lei 2338/2023, Brazil's risk-tiered "Marco Legal da IA," passed the Senate unanimously but sits in a Chamber special commission awaiting its rapporteur's report. Any company building or deploying AI in Brazil should treat a second, non-US compliance regime as coming.

By Anthony Guerriero · Reviewed by The Leveraged Years Editorial Desk · Published July 7, 2026 · Last updated July 7, 2026

Brazil's AI Act clears the Senate but stalls in the Chamber, final vote now expected in 2026 regulation briefing

The Leveraged Years AI Regulation Tracker

Regulatory Briefing

Bindingness: Non-Binding Framework
Scope: Brazil (national). All sectors and professions that develop or deploy artificial intelligence.
What Changed: PL 2338/2023, the "Marco Legal da Inteligencia Artificial," was approved unanimously by the Senate plenary in December 2024 after more than five years of processing. It is now before a Chamber of Deputies special commission that held 12 public hearings between May 20 and September 30, 2025, and is awaiting the rapporteur's report. The government-articulated text proposes a Conselho Brasileiro para Inteligencia Artificial coordinating a national AI governance system.
Who This Affects: Any company developing or deploying AI in Brazil, including US multinationals with Brazilian operations, and the compliance, legal and risk officers who advise them.
Current Status: Proposed. Approved by the Senate; pending in the Chamber of Deputies special commission.
Effective Date: Not set. Chamber floor vote expected in 2026 (date unconfirmed).
Primary Source: Camara dos Deputados, tramitacao of PL 2338/2023: https://www.camara.leg.br/proposicoesWeb/fichadetramitacao?idProposicao=2487262
Workflow Impact: Firms cannot yet comply with a final text, but should begin classifying AI use by risk and building the transparency and documentation practices the bill signals.
Operational Next Step: Track the special commission's tramitacao page and start an internal inventory of where AI touches decisions about people.
Last Verified: July 7, 2026

Brazil's attempt to pass a single, national law for artificial intelligence has cleared its hardest chamber and then stopped short of the finish line. Projeto de Lei 2338/2023, known as the Marco Legal da Inteligencia Artificial, was approved unanimously by the Senate plenary in December 2024 after more than five years of processing. It then moved to the Chamber of Deputies, where it now sits in a special commission awaiting the rapporteur's report. A plenary vote that was expected by the end of 2025 has slipped into 2026.

Where the bill actually stands

According to the Chamber's official tramitacao record, PL 2338/2023 is in the Comissao Especial created to issue an opinion on the bill, with the status "awaiting the rapporteur's opinion." The rapporteur is Deputy Aguinaldo Ribeiro, designated in May 2025, and the commission is presided over by Deputy Luisa Canziani. Between May 20 and September 30, 2025, the commission held 12 public hearings gathering input from industry, academia, rights holders and civil society. As of the latest verified record, no floor vote date has been scheduled.

That distinction matters. The Senate text is not the final law. In Brazil's process, the Chamber can amend the bill through a substitutivo, and any changes would send it back to the Senate. The consolidated substitutivo text and any scheduled floor date remain unverified, so professionals should treat the Senate version as a strong signal of direction rather than a settled rule.

What the bill would require

PL 2338/2023 is built as a risk-tiered, horizontal law. Rather than regulating one sector, it sets principles, rights and duties that scale with the risk a given AI system poses. The framework would impose obligations around transparency, risk assessment and accountability on organizations that develop or deploy AI. Higher-risk uses, the kind that affect access to services, employment or fundamental rights, would carry heavier documentation and governance duties.

A central open question is who firms would answer to. The government-articulated text proposes a Conselho Brasileiro para Inteligencia Artificial, or CBIA, to lead a Sistema Nacional para Desenvolvimento, Regulacao e Governanca de IA, referred to as the SIA. That structure would coordinate oversight across existing regulators. The alternative debated during the hearings would center governance on the existing data-protection authority. The resolution of that question determines the practical shape of compliance, from who issues guidance to who conducts enforcement.

What it does not do yet

For all its reach, PL 2338/2023 imposes no obligations today. There is no compliance deadline, no penalty in force, and no final statutory text to build against. A company cannot yet certify against the law because the law does not yet exist in final form. What it does provide is a clear preview of the duties Brazil intends to codify: know your AI systems, assess their risk, document that assessment, and be able to explain and account for automated decisions.

The cross-border angle

For a US reader, the significance is structural. Like the European Union's AI Act, Brazil is building a single, risk-tiered horizontal law rather than a sector-by-sector patchwork. A US multinational operating in Brazil would face a second non-US regime, distinct from any individual US state's rules, and would need a compliance approach that maps to Brazil's tiers. Firms that have already built EU AI Act programs will find the architecture familiar, which is the practical reason to start the internal inventory now rather than after the Chamber votes.

Frequently Asked Questions

What is PL 2338/2023 and what changed?

It is Brazil's proposed Marco Legal da Inteligencia Artificial, a risk-tiered national AI law. The Senate approved it unanimously in December 2024, and it is now pending in a Chamber of Deputies special commission awaiting the rapporteur's report, with a floor vote expected in 2026.

Who does the bill affect?

Any organization that develops or deploys AI in Brazil, including US companies with Brazilian operations. It would reach across sectors, with heavier duties for higher-risk uses that affect people's rights, employment or access to services.

Is the bill in force now?

No. It has passed the Senate but not the Chamber. There is no effective date, no penalty in force, and the final consolidated text is not settled. It remains a proposal.

What should professionals do while it is pending?

Start an inventory of where AI touches decisions, classify those uses by risk, and build transparency and risk-assessment documentation. Track the Chamber's tramitacao page for the rapporteur's report and any scheduled vote.