Brazil - National AI Policy (5.691/2019)

Overview

Projeto de Lei nº 5.691/2019 proposes the creation of a National Artificial Intelligence Policy in Brazil with the principal aim of stimulating an environment favorable to research, development and deployment of AI technologies. The bill is high-level and declaratory: it establishes principles (including ethics, human rights, privacy, transparency and security), directives to guide public policy and instruments such as transversal programs, sectoral funds and convênios for public–private cooperation. The original official bill text (avulso inicial) and committee reports are available from the Federal Senate; see the bill PDF and Senate matter page for the authoritative text and status. Senate PDF (initial text) and the Senate matter page track legislative steps and archival decision.

Definitions

The bill itself is concise and does not introduce extensive technical definitions for AI subcategories (e.g., machine learning, deep learning, foundation models). Its working definitions are normative: 'Artificial Intelligence' in the context of the Policy denotes technologies and solutions that enable automated reasoning, prediction or decision-making based on data processing. The text focuses on functional characteristics — autonomy, traceability, intelligibility and potential impacts — rather than detailed technical taxonomies. Implementing measures and subsequent regulatory initiatives would be expected to develop operational definitions for risk stratification, testing and oversight.

Governance and Institutional Framework

PL 5.691/2019 establishes an institutional orientation rather than a specific regulatory agency. The bill envisages cooperation among federal entities, public institutions and private sector actors and the use of convênios and funds to operationalize the policy. It therefore implies roles for ministries responsible for science, technology and innovation, for economic development, and for sectoral regulators, as well as coordination with data protection authorities in relation to privacy safeguards. For official procedural records and the list of committee rapporteurs and actions, consult the Senate matter page and reports. Senate matter page provides the legislative timeline and committee assignments.

Key Focus Areas

PL 5.691/2019 emphasizes multiple policy priorities. (1) Ethical and rights-respecting AI: respect for human rights, democratic values and social diversity. (2) Privacy and data protection: protection of personal data and private life consistent with LGPD expectations. (3) Transparency and explainability: intelligibility, traceability of decisions and democratic scrutiny. (4) Security and reliability: safety mechanisms, human intervention and risk mitigation. (5) Economic and social objectives: promotion of inclusive and sustainable growth, R&D stimulation, incentives for private investment, capacity building and workforce development. (6) Institutional cooperation: incentives for partnerships, international collaboration and cross-sectoral programs. The bill lists concrete instruments to pursue these aims, including sectoral science and innovation funds and convênios between public and private entities.

Implementation Framework

The bill prescribes instruments (Article 5) such as transversal programs in partnership with public bodies and private institutions, sectoral science and technology funds and convênios for developing social technologies. Article 6 explicitly authorizes the Union and public entities to celebrate convênios with national or international partners to obtain technical, human or financial resources to support the Policy. The bill does not include implementing regulations or detailed technical standards (testing, conformity, certification); it leaves the operationalization of standards, risk-assessment methodologies and enforcement modalities to subsequent executive action, implementing regulations and sectoral authorities.

Monitoring and Evaluation

PL 5.691/2019 does not set an elaborate monitoring or reporting regime within its text. Monitoring and evaluation are implied through institutional cooperation and programmatic instruments (funds and convênios), and would likely be implemented via reporting requirements attached to specific programs, funds and agreements. Committee reports and later consolidated substitute texts in 2024 provide more detailed oversight proposals; the bill itself contemplates continuous governance standards and risk mitigation obligations as normative expectations for AI solutions.

Penalties, Liability, and Appeals

The bill is silent on administrative penalties, criminal sanctions or a specific liability regime. PL 5.691/2019 focuses on principles, directives and instruments, leaving concrete enforcement, liability frameworks and appeals processes to be developed through subsequent legislation, regulations or sector-specific rule-making. As such, stakeholders seeking precise compliance obligations must look to implementing acts and other laws (for example, LGPD for data-protection liability) and to later consolidated AI bills and committee substitutes.

Relationship to Other Instruments

PL 5.691/2019 should be read alongside Brazil’s Lei Geral de Proteção de Dados (LGPD, Law No. 13.709/2018) on personal data protection, sectoral regulatory regimes (e.g., health, financial services) and other AI-oriented bills that were processed together in the Senate’s temporary commission (e.g., PL 5051/2019, PL 2338/2023, PL 21/2020). The Senate’s committee deliberations in 2023–2024 consolidated many AI proposals; that consolidation led to the adoption of substitutes which ultimately caused PL 5.691/2019 to be declared prejudiced and archived on 10 December 2024. Refer to the Congresso Nacional packet for bicameral tramitation records. Congressional matter page.

International Alignment

The bill affirms international collaboration as a principle and directive, encouraging cooperation with foreign institutions to exchange information, foster collaboration and obtain technical resources. This orientation aligns PL 5.691/2019 with global policy trends emphasizing ethics, human rights and cross-border research cooperation. The bill does not, however, codify alignment with specific international instruments (e.g., OECD AI Principles or European Union AI Act) but it sets a permissive framework for future harmonization through convênios and cooperation agreements.

Implementation Timeline

Milestone	Date
Bill introduced in Senate	2019-10-24
Reading in Senate plenary	2019-10-25
Legislative report (CCT)	2022-04-13
Considered by CTIA and consolidated with other AI bills	2023-2024
Laid as prejudiced / archived	2024-12-10

Compliance Checklist

Action	Notes
Map AI uses and data flows	Ensure alignment with privacy and intelligibility principles
Adopt human-in-the-loop safety measures	Design interventions and escalation points
Document model decisions and traceability	Enable democratic scrutiny and auditability
Engage in public–private convênios	Pursue funding and partnerships consistent with policy instruments
Design capacity-building programs	Focus on workforce transition and reskilling

Sources and References

Source	Type
SENADO FEDERAL — PL 5691/2019 (avulso inicial PDF)	Primary Source
Senado Federal — Matter page for Projeto de Lei nº 5691/2019	Primary Source
Congresso Nacional — Bicameral matter page (PL 5691/2019)	Primary Source