In the age of AI’s omnipresence, Brazil is making significant strides to foster a robust ecosystem for the responsible development of artificial intelligence through Brazilian AI Bill. The widespread adoption of AI offers unprecedented economic opportunities and, simultaneously, poses unique challenges, particularly in the realms of workforce transformation and algorithmic biases. To address this rapidly evolving landscape, Brazil is working to enhance its digital infrastructure to encourage AI innovation while implementing essential regulatory measures.
Brazil is actively shaping national guidelines and initiatives aimed at bolstering the necessary human and physical infrastructures, while also participating in international discussions on best practices in AI. As AI continues to reshape society on a fundamental level, Brazil, representing the Global South, is contributing its unique perspectives to foster a human-centered, inclusive, development-oriented, responsible, and ethical approach to AI. The overarching goal is to enhance people’s lives and bridge the digital divide.
Table of Contents
The Brazilian National Strategy for Artificial Intelligence (EBIA)
In April 2021, the Brazilian government introduced the National Strategy for Artificial Intelligence, known as EBIA, leveraging the nation’s strengths in technology, innovation, and entrepreneurship to address its challenges and create new opportunities within the innovation ecosystem. EBIA aligns itself with the five principles defined by the OECD AI Principles:
- Inclusive growth, sustainable development, and well-being.
- Values centered on human beings and equity.
- Transparency and explainability.
- Robustness, security, and protection.
The Brazilian government is committed to ensuring that the benefits of AI are distributed equitably and sustainably throughout society. To achieve this, various AI initiatives, supported by the Brazilian government, are divided into nine strategic axes:
- Legislation, regulation, and ethical use.
- AI governance.
- International aspects.
- Qualifications for a digital future (Education).
- Workforce and training.
- Research, Development, Innovation, and Entrepreneurship.
- Application in productive sectors.
- Application in the public sector.
- Public security.
While the first three axes are transversal, applying to all other areas, the remaining six are vertical, focusing on sector-specific issues. The EBIA encompasses a total of 73 strategic actions.
EBIA: Significant Achievements and Initiatives
Since the launch of EBIA, Brazil has made noteworthy progress. The country established six applied centers for AI (CPA) in areas such as smart cities, agriculture, Industry 4.0, and health. Existing AI-focused institutions, including the Center for AI (C4AI) and the Brazilian Association of Research and Industrial Innovation Network of Digital Technologies and Innovation (Embrapii’s Network), have further expanded their efforts.
Another substantial initiative is EMBRAPII’s CIMATEC Lab, the first infrastructure in Latin America dedicated to industry innovation. It boasts five supercomputers with a processing capacity of 4 PFLOPS and was inaugurated in 2021.
Additionally, Brazil has implemented grants for startups and education programs, including “The Future of Work, Work of the Future,” aimed at preparing the current workforce across all education levels in fields such as data science, cybersecurity, the Internet of Things (IoT), cloud computing, and robotics.
With these initiatives in motion, Brazil is reinforcing its position in AI technology to tackle national challenges by enhancing the skills of its workforce and capitalizing on AI-driven transformations. To make this vision a reality, Brazil has committed to regulating AI and participating in global cooperation to promote trustworthy AI principles, exemplified by its involvement in the Global Partnership on Artificial Intelligence (GPAI) and other international forums.
The Brazilian AI Bill
Brazil has been actively exploring the regulation of AI systems since 2019. Various bills, including nº 5.051/2019, nº 872/2021, and nº 21/2020, were introduced in the Brazilian National Congress. However, public debate on these bills was limited due to the COVID-19 pandemic.
In early 2022, the Chamber of Deputies approved Bill nº 21/2020, which then moved to the Federal Senate. To further shape AI legislation, the Higher House formed a Commission of Legal Experts, known as CJSUBIA, comprising 18 members with expertise in technology law and regulation.
CJSUBIA conducted a series of public hearings in April and May 2022, featuring over 50 specialists from diverse groups, including public authorities, the business sector, civil society, and the scientific-academic community. The commission’s work emphasized gender parity and regional diversity and revolved around four key themes: (i) concepts and classification of artificial intelligence, (ii) AI’s impacts, (iii) rights and duties, and (iv) accountability, governance, and supervision.
In June 2022, CJSUBIA organized an international seminar to gather global expert opinions on these topics. Over six months, the group researched similar legislations and initiatives in 31 OECD countries, leading to a comprehensive 900-page report published in December 2022, including a draft of an alternative bill.
This report culminated in the proposal of Bill nº 2.338/2023 in May 2023, by the President of the Federal Senate, replacing the former Bill nº 21/2020. The new bill will be presented to the Science, Technology and Innovation Committee of the Federal Senate for approval, and if passed, it will proceed to the Plenary of the Higher House and, ultimately, to the Chamber of Deputies.
Key Highlights of Brazilian AI Bill: Bill nº 2.338/2023:
- A Human Rights-Oriented Approach: The bill proposes rights for individuals affected by AI systems, encompassing rights to preliminary information, explanations of AI-generated decisions, non-discrimination, and privacy protection.
- Risk Classification: The bill introduces three risk levels for AI systems – excessive risk, high risk, and non-high risk. High-risk systems must undergo a preliminary self-assessment analysis by the AI provider before deployment.
- AI Governance: All AI systems, regardless of their risk level, must implement governance structures, including transparency and security measures. High-risk systems must also provide technical documentation, log registers, reliability tests, bias mitigation, and technical explainability measures.
- Supervisory Authority: The bill assigns powers to a supervisory authority to regulate and enforce AI legislation, with the specific body designated by the Executive Branch. This authority is expected to coordinate its activities with other sectoral authorities regulating AI.
- Other Provisions: The bill includes rules for civil liability, codes of best practices, notification of AI incidents, administrative sanctions, fostering innovation through regulatory sandboxes, and creating copyright exceptions for data mining processes. It also proposes an open public database of high-risk AI systems held by the AI supervisory authority.
The OECD and International Alignment
While the final form of Brazil’s AI legislation is yet to be determined, the ongoing efforts in the Brazilian Congress and other public bodies implementing the National AI Strategy align closely with the OECD Recommendations and other international guidelines. These initiatives are paving the way for Brazil’s development and deployment of responsible and trustworthy AI.