Senate Committee approves proposal to regulate AI in Brazil

This Thursday (5), the temporary Senate committee created with the purpose of analyzing Bill 2,338/2023, which regulates artificial intelligence (AI), approved a proposal that provides for the rules for using the technology in Brazil.

The next stage will take place next Tuesday (10), when there will be a vote in the main plenary of the legislative house. The text of the proposal was reported by Senator Eduardo Gomes (PL) and will only become Law after approval in the Senate and, subsequently, in the Chamber of Deputies. If approved by both legislative houses, the project will go to the President of the Republic Luiz Inácio Lula da Silva for sanction.

What does the proposal say

The project establishes principles to protect democracy, the integrity of information and vulnerable groups. The rules will take into account the possible harms of technology established to fundamental, individual and social rights established in the Federal Constitution. Mechanisms will also be adopted against any types of discrimination and violence.

Artificial intelligence tools must be identified using symbols. The proposal ensures the copyright of third parties and, to this end, developers must duly remunerate the authors of the content.

The proposal establishes limits on the exploration and development of AI systems. Furthermore, it sets a precedent for possible judicial punishments with specific definitions for penalties

National Artificial Intelligence Regulation and Governance System

Coordinated by the National Data Protection Authority (ANPD), the National Artificial Intelligence Regulation and Governance System will be the body responsible for supervising artificial intelligence systems in Brazil.

The regulatory body will have the role of creating a database and public data system containing impact assessments and possible damages caused by AI.

_{If approved by the Senate and the Chamber, the project will go to the President of the Republic for sanction (Video: reproduction/YouTube/@TVSenado)}

What is prohibited

The project foresees prohibitions on the use of AI in Brazil. Prohibited conduct includes: producing and disseminating content with sexual exploitation and abuse of minors; exploit people in vulnerable situations; induce data to security, fundamental rights and health; expose and evaluate psychological personality traits; among others.

AI companies must also identify copyright protection seals on their content. Furthermore, those responsible for production, such as journalists, artists and companies, will have their content protected by copyright. In cases of generating new materials, the tools must inform channels so that the authenticity of the information can be verified.

Punishments

If the rules are broken, there will be punishment, civil liability for damages to third parties and administrative punishment. Fines could reach R$50 million or 2% of the company's revenue. In addition, restrictions may occur such as partial or total, temporary or definitive suspension of the platform, among others.

Featured photo: President of the Council Miguel Matos and senator Eduardo Gomes, during a meeting (Reproduction: Edilson Rodrigues/Agência Senado)