Artificial Intelligence (AI) is already an integral part of key sectors such as healthcare, transportation, finance, and services. Given its transformative potential — but also the risks associated with unchecked use — Brazil is moving forward with a robust regulatory proposal: Bill No. 2338/2023, currently under review in the Chamber of Deputies.
The bill’s core aim is to establish clear rules and fundamental principles for the development and use of AI in the country. It is guided by values such as human dignity, privacy, equality, and transparency, encouraging responsible and ethical innovation. It also seeks to ensure essential user rights, including the explainability of automated decisions, the possibility of contesting them, and continuous human oversight in sensitive processes.
A central feature of the bill is the risk-based classification of AI systems. Two main categories are proposed:
- Excessive risk, which includes uses that are outright prohibited — such as systems that exploit individual vulnerabilities, manipulate behavior subliminally, or enforce unfair social scoring mechanisms, particularly in governmental contexts.
- High risk, covering applications with direct impact on fundamental rights in areas such as health, justice, security, credit, employment, education, critical infrastructure, autonomous vehicles, and biometric identification. For these, the bill mandates stringent governance measures, including public algorithmic impact assessments, safety testing, and human oversight mechanisms.
Another key pillar of the proposal is governance and accountability. All AI systems, regardless of risk level, must have clear technical documentation, identifiable responsible parties, understandable decision explanations, and strategies to prevent algorithmic discrimination. In high-risk cases, civil liability will be strict — meaning companies can be held liable without proof of fault — with reversal of the burden of proof in favor of the affected party.
As for penalties, the bill establishes fines of up to R$ 50 million or 2% of the infringing group’s revenue, in addition to other sanctions such as warnings, suspension, or permanent bans on system use.
The proposal also provides for the creation of the National AI Regulation and Governance System (SIA), coordinated by Brazil’s National Data Protection Authority (ANPD), and involving regulators, experts, and civil society representatives. This body will be responsible for overseeing implementation, periodically reviewing the rules, and harmonizing their application across industries.
As of July 2025, Bill 2338/2023 is under review by a special committee in the Brazilian Chamber of Deputies. Already approved by the Senate, it remains a legislative priority, with ongoing public discussions. Its final approval could represent a turning point in the regulation of artificial intelligence in Brazil.
For companies developing or using AI technologies, this bill represents a strategic milestone: it offers greater legal certainty, aligns Brazil with global regulatory trends (especially the EU’s AI Act), and lays the foundation for a safe and responsible digital transformation.
At GT Lawyers, we are closely monitoring the developments of PL 2338/2023 and are ready to support your company with regulatory compliance, AI governance best practices, and legal risk mitigation strategies.
