English 
Spanish 
5 min read
0 comments
Artificial Intelligence PL slows down innovation in Brazil with excessive regulation, imposing barriers to technological development. Industry defends focus on risks.
The Artificial Intelligence bill threatens innovation in Brazil due to overly strict regulation, creating obstacles to technological growth. The industry says the focus should be on specific risks. Technology companies express concern about the impact on competitiveness, considering the proposed restrictions.
The proposed legislation for AI, known as PL 2.338/2023, is seen as one of the most restrictive globally and could compromise Brazil's position on the international stage. The productive sector highlights that such regulations could stifle innovation and negatively affect economic progress. The industry calls for regulation to be balanced, mitigating risks without hindering the significant advances provided by Artificial Intelligence.
Regulatory Focus on AI by Risk
Regulation should focus on the use of Artificial Intelligence (AI) based on risks, not technological development of the technology itself. The industry supports regulation of use of AI in Brazil, but warns about the barriers created by Bill of Law 2.338/2023, which may hinder technological development and innovation in the country. The text, presented on Thursday (4) at the Temporary Commission on Artificial Intelligence in Brazil (CTIA) of the Federal Senate, brings specific advances in relation to previous versions. However, it maintains a structure and conceptual basis that result in a regulatory model of great scope and rigor, the most severe in the world, according to experts, which could lead the country to face isolation and technological delay.
- Charges to 100% in 10 minutes! New technology enables lithium-ion batteries to achieve ultra-fast charging speeds for electric vehicles
- Xiaomi revolutionizes the market with a new air conditioner that cleans itself, is up to 3x quieter, has advanced technology and is here to be the market leader
- Scientists didn't know what to do with this toxic waste, but they discovered that it is a treasure trove of rare metals valued at $8,4 billion
- Startup Plans to Refreeze the Arctic with Underwater Drones for an Estimated $10 Billion a Year
See more
Position of industrial leaders
Brazilian industrial leaders, gathered at the Business Mobilization for Innovation (MEI) and the National Confederation of Industry (CNI), recognize the importance of preventing the misuse of AI, especially in the context of elections and social networks. The CNI highlights that the Bill exceeds its scope by regulating AI from the conception and development of systems, instead of focusing on use and applications based on the degree of risk, as is done in other countries. Jefferson Gomes, director of Technology and Innovation at CNI, emphasizes that Brazilian industry is deeply integrated into value chain of AI, providing inputs and developing infrastructure fundamental resources such as energy, hardware, chips and operating systems, in addition to being an application developer and pioneering user, especially in the agro and services sectors.
Impacts of regulation on the development of AI
Brazil has competitive advantages and AI has great potential to boost productivity in the national industry. However, Jefferson Gomes warns that the proposed regulation ignores the different uses and associated risks, imposing a heavy regulatory apparatus not only on AI applications, but also on the research and development of the technology. This could scare away new investments, harm AI projects in the productive sector that do not use personal data and, ultimately, lead the country to lose competitiveness and the opportunity to consolidate itself as an important player in global chains.
Critical points of the bill
The most critical points of the text of Pl 2.338/2023 include:
1) Regulation from conception to adoption of AI systems, and not just the use and implementation of applications based on risk, which impedes technological development;
2) The creation of a letter of rights dissociated from risk analysis, instead of defining obligations of the regulated agent;
3) An excessive governance burden, even for low-risk applications, in addition to external intervention in companies' internal processes, which violates commercial and industrial secrets;
4) Overlay regulatory powers between central bodies and a central authority, generating legal uncertainty;
5) The choice of the National Data Protection Authority (ANPD) as the central body, which is still in the structuring phase and does not have technical competence on AI or innovation policies like other government bodies.
Industry Proposals for AI Regulation
The industry proposes several improvements to AI regulation, including:
– Scope adjustment to focus on system applications, where it is possible to materialize and assess risks;
– Regulation by risk, differentiating high-risk applications from low- and medium-risk applications, especially those that do not involve personal data or human interaction;
– Adequacy of the regulatory model to define the regulated agent’s obligations towards the regulator, and not the user’s rights over the service provider;
– Expand the decentralization and the role of sectoral agencies in the National Artificial Intelligence Regulation and Governance System (SIA);
– Adequacy of the regulatory proposal and government policies to stimulate innovation, such as the Brazilian Artificial Intelligence Strategy (EBIA) and the Industrial Policy Action Plan (PBIA).
Brazil's Competitive Differentiators
Brazil has important competitive advantages for the development and use of AI, such as the size and heterogeneity of the population, which provide valuable data to train generative AI applications, in addition to a energy matrix clean, able to respond to the demands of datacenters. Global movements aim to decarbonize production processes, directing production to regions with clean, safe and abundant energy, exemplifying the movement known as powershoring.
AI Regulation in Other Countries
Comparing with other countries:
– In the European Union, AI regulation is risk-centric, restricting more severely only high-risk applications while allowing greater flexibility for limited-risk ones;
– In the UK, the emphasis is on the benefits of AI and incremental regulation, focusing on use rather than technology;
– The United States avoids regulatory overreach, providing a culture of self-regulation;
– Japan and Singapore follow similar practices, with certifications to ensure good practices in AI, updated according to technological developments.
These international practices offer a broad understanding of possible paths to effectively regulating AI, balancing innovation and security.
Source: Journalism – CNI
