The Brazilian Senate committee released a report on December 1, 2022, regarding the regulation of artificial intelligence (AI). This report includes a proposed draft law for AI regulation, presented in Portuguese, which will serve as the foundation for Senate discussions on new AI legislation. The committee's work involved drawing insights from previous Brazilian AI regulation proposals, studying how OECD nations are approaching AI regulation, and considering inputs from public hearings and stakeholder feedback. Brazil's Draft AI Law establishes a comprehensive framework for AI development and usage. Guided by key principles such as transparency and non-discrimination, the law defines AI systems and mandates risk assessments for providers and users. High-risk AI systems are subject to stricter regulations, including the maintenance of a public database for them. The law prohibits harmful AI techniques and protects user rights, with governance structures and security measures required. Liability is imposed on providers and users for AI system damages, with copyright considerations. Sandboxes support AI innovation, and enforcement is overseen by a competent authority with potential fines. The law aims to balance innovation and ethical AI use in Brazil.
