Non-official Translation of the Brazilian Artificial Intelligence Bill, n. 21/2020

This is a non-official translation of the bill establishing foundations, principles and guidelines for AI in Brazil prepared by the CyberBRICS team – translated by Walter B. Gaspar, Eduardo Mattos and Luca Belli. The original text in Portuguese, alongside more information on the bill and its processing in the Brazilian Congress, can be found at the Brazilian Chamber of Deputies’ website. The bill was approved in Congress and currently awaits a vote in the Senate.

Final text, Bill n. 21-A/2020

Establishes foundations, principles and guidelines for the development and application of artificial intelligence in Brazil, and gives further provisions.


Art. 1. This Law establishes foundations and principles for the development and application of artificial intelligence in Brazil and provides guidelines for incentives and activities of public entities regarding the subject.

Art. 2. For the purposes of this Law, an artificial intelligence system is defined as a system based on a computational process which, from a selection of objectives determined by humans, may, through data and information processing, learn to perceive, interpret and interact with the external environment, making predictions, recommendations, classifications or decisions, and utilizing techniques such as, but not limited to, the following:

I – machine learning systems, including supervised, unsupervised and reinforcement learning;

II – logic or knowledge based systems;

III – statistical approaches, bayesian inference, methods of research and optimization.

Sole paragraph. This Law does not apply to automation processes exclusively oriented by pre-defined programming parameters that do not include the system’s capability to learn and perceive, interpret and interact with the external environment from the actions and information received.

Art. 3. The application of artificial intelligence in Brazil is aimed at scientific and technological development, as well as:

I – incentivizing sustainable and inclusive economic development and societal welfare;

II – increasing Brazilian competitiveness and productivity;

III – incorporating Brazil into the global value chains in a competitive position;

IV – improving public services and the implementation of public policies; and

V – promoting research and development with the purpose of stimulating innovation in the productive sectors; and

VI – protecting and preserving the environment.

Art. 4. Development and application of artificial intelligence in Brazil have the following foundations:

I – scientific and technological development and innovation;

II – free enterprise and competition;

III – observance of ethics, human rights and democratic values;

IV – freedom of thought and freedom of expression of intellectual, artistic, scientific and communicative activity;

V – non-discrimination, plurality, respect for regional diversity, inclusion and respect for the fundamental rights and guarantees of citizens;

VI – acknowledgement of its digital, transversal and dynamic nature;

VII – stimulus to self-regulation, through adoption of conduct codes and good practice guidelines, observing the principles enumerated in art. 5th and global good practices;

VIII – security, privacy and personal data protection;

IX – information security;

X – access to information;

XI – national defense, State security and national sovereignty;

XII – freedom of business models, barring those conflicting with the determinations of this Law;

XIII – preservation of stability, security, resiliency and functionality of artificial intelligence systems, through technical measures compatible with international standards and through incentives to the adoption of good practices;

XIV – protection of free competition and against abusive market practices, as stated in Law n. 12.529, of November 30th, 2011; and

XV – harmonization with Laws n. 13.709, of August 14th, 2018 (General Data Protection Law, “LGPD”), n. 12.965, of April 23rd, 2014 (Civil Framework for the Internet, “MCI”), n. 8.078, of September 11, 1990 (Consumer Protection Code, “CDC”), and n. 12.527, of November 18, 2011 (Access to Information Law, “LAI”).

Sole paragraph. Conduct codes and good practices guidelines mentioned in item VII may serve as indicators of compliance.

Art. 5. The following are the principles for the development and application of artificial intelligence in Brazil:

I – beneficial purpose: artificial intelligence systems shall seek beneficial results for humanity;

II – centrality of the human being: respect for human dignity, privacy, personal data protection and fundamental rights, when the system deals with matters related to the human being;

III – non-discrimination: mitigate the possibility of using systems for illicit or abusive discriminatory purposes;

IV – pursuit of neutrality: it is recommended that the agents involved in the development and operation of artificial intelligence systems strive to identify and mitigate biases that are contrary to the determinations of current legislation;

V – transparency: people have the right to be informed in a clear, accessible and precise manner about the use of artificial intelligence solutions, unless otherwise provided by law, and subject to commercial and industrial secrets, in the following cases:

a) when they are directly communicating with artificial intelligence systems, such as through conversation robots for personalized online services (chatbot), while using these systems;

b) on the identity of the natural person, when they operate the system autonomously and individually, or of the legal entity responsible for the operation of artificial intelligence systems;

c) on the general criteria that guide the functioning of the artificial intelligence system, observing commercial and industrial secrets, whenever there is a potential of considerable risk to fundamental rights;

VI – safety and prevention: use of technical, organizational and administrative measures, considering the use of means that are reasonable and available at the time, compatible with best practices, international standards and economic feasibility, aimed at allowing the management and mitigation of risks arising from the operation of artificial intelligence systems throughout their life cycle and their continuous operation;

VII – responsible innovation: agents who work in the development and operation of artificial intelligence systems in use must ensure the adoption of the provisions of this Law, documenting their internal management process and being responsible, within the limits of their respective participation, considering the context and available technologies, for the results of the functioning of these systems;

VIII – availability of data: the use of data, databases and texts protected by copyright for the purposes of training artificial intelligence systems does not imply the violation of these rights, as long as it does not impact the normal exploitation of the work by its owner;

Art. 6. When regulating the application of artificial intelligence, the government must observe the following guidelines:

I – subsidiary intervention: specific rules shall be developed for the use of artificial intelligence systems only when absolutely necessary to guarantee compliance with the provisions of current legislation;

II – sectoral intervention: intervention by public entities must be conducted by the competent body or entity, considering the context and regulatory framework specific to each sector;

III – risk-based management: the development and use of artificial intelligence systems should consider the concrete risks and definitions on the need for regulation of artificial intelligence systems and the respective degree of intervention should always be proportional to the concrete risks offered by each system and the probability of occurrence of these risks, always evaluated in reference to:

a) the potential social and economic benefits offered by that artificial intelligence system; and

b) the risks presented by similar systems that do not involve artificial intelligence, pursuant to item V;

IV – social and interdisciplinary participation: the adoption of norms that impact the development and operation of artificial intelligence systems will be based on evidence and preceded by public consultation, preferably carried out via the internet and with wide prior dissemination in order to enable the participation of all the interested parties and the various specialties involved;

V – regulatory impact analysis: the adoption of standards that impact the development and operation of artificial intelligence systems will be preceded by regulatory impact analysis, pursuant to Decree n. 10.411, of 2020 and Law n. 13.874, of 2019; and

VI – liability: rules on the liability of agents who work in the chain of development and operation of artificial intelligence systems must, unless otherwise stipulated by law, be based on subjective liability, take into account the effective participation of these agents, the specific damages to avoid or remedy, and how these agents can demonstrate compliance with applicable standards through reasonable efforts consistent with international standards and best market practices.

§1 Regarding the risk-based management presented in item III above, in cases of low risk, the public administration shall encourage responsible innovation with the use of flexible regulatory techniques.

§2 Regarding the risk-based management presented in item III above, in specific cases where a high risk is found, the public administration may, within its competence, request information on the safety and prevention measures listed in item VI of article 5, and respective safeguards, under the terms and limitations of transparency established by this law, observing commercial and industrial secrets.

§3 When the use of the artificial intelligence system involves consumer relations, the agent is liable regardless of guilt for the reparation of damages caused to consumers, within the limit of their effective participation in the harmful event, in compliance with Law No. 8.078 of September 11, 1990.

§4 Legal entities under public law and those under private law that provide public services shall be liable for damages that their agents, in this capacity, cause to third parties, with the right of recourse against the person responsible in cases of intent or negligence.

Art. 7. The following are guidelines for the performance of the Union, States, Federal District and Municipalities in relation to the use and promotion of artificial intelligence systems in Brazil:

I – promote trust in artificial intelligence technologies, disseminating information and knowledge about their ethical and responsible uses;

II – incentivize investments in research and development of artificial intelligence;

III – promote technological interoperability of artificial intelligence systems used by the government, in order to allow for the exchange of information and speeding up of procedures;

IV – encourage the development and adoption of artificial intelligence systems in the public and private sectors;

V – encourage training and preparation of people for the restructuring of the labor market;

VI – encourage innovative pedagogical practices, with a multidisciplinary vision, and emphasize the importance of reframing teacher education processes to deal with the challenges arising from the insertion of artificial intelligence as a pedagogical tool in the classroom;

VII – encourage the adoption of regulatory instruments that promote innovation, such as experimental regulatory environments (regulatory sandboxes), regulatory impact analysis and sectorial self-regulation;

VIII – encourage the creation of transparent and collaborative governance mechanisms, with the participation of representatives of the government, the business sector, civil society and the scientific community; and

IX – promote international cooperation, stimulating the sharing of knowledge about artificial intelligence systems and the negotiation of treaties, agreements and global technical standards that facilitate interoperability between the systems and the harmonization of legislation in this regard.

Sole paragraph. For the purposes of this article, the Federal Public Administration will promote strategic management and guidance on the transparent and ethical use of artificial intelligence systems in the public sector, in accordance with strategic public policies for the sector.

Art. 8. The guidelines referred to in articles 6 and 7 shall be applied in accordance with the regulations of the Federal Executive Branch by sectoral bodies and entities with technical competence in the matter, which shall:

I – monitor the risk management of artificial intelligence systems, in specific cases, evaluating the risks of applications and the mitigation measures in their area of competence;

II – establish rights, duties and responsibilities; and

III – recognize self-regulatory institutions.

Article 9. For the purposes of this Law, artificial intelligence systems are technological representations from the field of information technology and computer science, and it is the exclusive competence of the Union to legislate and regulate the matter to promote legal uniformity throughout national territory, in the form of the provisions of art. 22, IV of the Federal Constitution.

Art. 10. This Law enters into force ninety days after the date of its official publication.

Sessions Room, September 29, 2021.

Representative LUISA CANZIANI


1 thought on “Non-official Translation of the Brazilian Artificial Intelligence Bill, n. 21/2020”

Comments are closed.