Artificial Intelligence · Brazil
AI regulation in Brazil: laws & policy (2026)
Brazil shaded by its artificial intelligence status
AI in Brazil: proposed.
FrameworkBill PL 2338/2023 ('Marco Legal da Inteligência Artificial') — approved by the Federal Senate in December 2024, pending final vote in the Chamber of Deputies (rescheduled to February 2026). No comprehensive AI law is yet in force; the National Data Protection Authority (ANPD) is designated as the future coordinator of the proposed National System for the Regulation and Governance of AI (SIA). In parallel, Brazil has a non-binding national strategy (EBIA, Portaria MCTI 4.617/2021) and the Plano Brasileiro de Inteligência Artificial (PBIA) 2024–2028.
Brazil does not yet have a comprehensive, in-force AI law. Its flagship risk-based bill (PL 2338/2023) passed the Senate on 10 December 2024 and is awaiting final approval by the Chamber of Deputies, with the vote pushed to early 2026 amid political disputes. AI governance today therefore rests on a patchwork of the LGPD (data protection), sectoral rules (e.g., ANPD, Central Bank, Anatel, Anvisa), the non-binding EBIA strategy, and the PBIA 2024–2028 investment plan, while the ANPD has flagged AI as a top enforcement priority for 2026–2027.
Key points
The Senate approved PL 2338/2023 on 10 December 2024 and sent it to the Chamber of Deputies. The Special Committee reported the text but the plenary vote has been postponed to February 2026, so the 'Marco Legal da IA' is not yet law.
The bill classifies AI into excessive-risk (prohibited, e.g., government social scoring, subliminal manipulation), high-risk (regulated, e.g., critical infrastructure, health, credit, hiring, biometric ID) and other systems, with obligations for algorithmic impact assessments, human oversight, transparency and documentation for high-risk systems.
PL 2338/2023 sets up the Sistema Nacional de Regulação e Governança da IA (SIA), a networked framework in which the ANPD is the coordinating authority and residual regulator, alongside sectoral bodies (Anatel, Anvisa, Anac, Aneel, Bacen, etc.). Fines can reach BRL 50 million per infraction or 2% of Brazilian turnover.
Even without a dedicated AI law, the ANPD supervises AI-related personal data processing under the LGPD (Law 13.709/2018). Its 2026–2027 Priority Themes Map (Dec. 2025) and the updated 2025–2026 Regulatory Agenda list AI and emerging technologies — including facial recognition, recommender systems and generative AI — as top enforcement priorities. Provisional Measure 1.317/2025 turned the ANPD into an independent regulatory agency with stronger enforcement powers.
The Estratégia Brasileira de Inteligência Artificial (EBIA) was instituted by Portaria MCTI No. 4.617 of 6 April 2021 (amended by Portaria 4.979/2021). It is a non-binding policy document with 3 transversal and 6 vertical axes covering ethics, R&D, workforce, application in productive sectors and public administration.
In July 2024 the Lula government launched the Plano Brasileiro de Inteligência Artificial 2024–2028 ('IA para o Bem de Todos'), a BRL 23 billion investment plan structured around five axes (infrastructure, workforce, applied research, business innovation and public services) coordinated by the MCTI and the Comitê Interministerial de IA.
Timeline - major decisions & events
Brazil's data protection authority published the final results of Notice No. 2/2025, choosing three companies for its first AI-and-data-protection regulatory sandbox focused on algorithmic transparency and human oversight. It marks the regulator's move from study to hands-on supervisory experimentation with generative AI.
Digital Policy Alert ↗Because PL 2338/2023 was referred to more than three standing committees, the Chamber's president established a special committee to analyze the AI framework. The move set the stage for contentious debate over biometric-surveillance carve-outs and copyright rules.
Library of Congress ↗The ANPD's Technology and Research Unit released a Technological Radar study on generative AI, clarifying how existing LGPD rules apply to model training and deployment. It signaled the regulator would govern AI through data-protection law pending a dedicated statute.
Future of Privacy Forum ↗After Meta agreed to safeguards, pseudonymization, improved opt-out, and a commitment not to use minors' data, the ANPD lifted its preventive measure. The compliance plan set a de facto template for how AI training on Brazilians' personal data must be conducted.
ANPD ↗President Lula received the first Plano Brasileiro de Inteligência Artificial, committing R$23 billion through 2028 for AI infrastructure (including the Santos Dumont supercomputer), public services, and sovereignty. It is Brazil's flagship industrial-policy push on AI.
Agência Gov (EBC) ↗The ANPD issued a preventive measure suspending Meta's updated privacy policy that used Facebook, Instagram and Messenger data to train its AI, citing improper legal basis and risks to minors, with a R$50,000 daily fine. It was Brazil's first major AI enforcement action.
ANPD ↗Drawing on the work of a Senate commission of jurists (CJSUBIA), the Senate president introduced PL 2338/2023 to replace the earlier Chamber bill with a rights-based, risk-tiered framework. This became the backbone of Brazil's current AI legislation.
Data Privacy Brasil ↗A committee of jurists was formed in the Senate to consolidate competing bills (PLs 5051/2019, 21/2020, 872/2021) and draft a substitute, holding public hearings with 50+ experts. Its work shifted Brazil away from the lighter-touch Chamber approach.
Data Privacy Brasil ↗The Chamber of Deputies passed the Marco Legal da Inteligência Artificial by 413-15, a principles-based bill criticized for weak protection of fundamental rights. Sent to the Senate, it was ultimately superseded by PL 2338/2023.
Câmara dos Deputados ↗The Ministry of Science, Technology and Innovation released the Estratégia Brasileira de Inteligência Artificial, with nine thematic axes and 73 strategic actions on ethics, governance, research and skills. It was Brazil's first government-wide AI policy roadmap.
MCTI ↗Law No. 13,709 created Brazil's comprehensive data-protection regime and the basis for the ANPD, in force from September 2020. Until a dedicated AI statute passes, the LGPD is the principal legal tool governing AI systems that process personal data.
Library of Congress ↗Brazil - other topics
Artificial Intelligence in other countries
Last verified 8/13/2026 · Orientation, not legal advice - verify against the primary sources linked above. Methodology & how to cite · State of Technology Regulation 2026 · Explore the full world map →