On August 1st, the AI Act, (EU) European Union Artificial Intelligence Act, proposed by the European Parliament came into effect, aiming to regulate artificial intelligence (AI) systems. Although it is a regulation of the European Union (EU), it has extraterritorial effects, applying to any AI system that impacts the market or citizens from the EU, regardless of whether it operates within European territory. Furthermore, considering the need for a coordinated and global approach to addressing AI, the EU AI Act will undoubtedly serve as inspiration for future regulations worldwide, including the one in Brazil.
The regulation classifies AI practices into risk categories and prohibits those that pose unacceptable risks. This includes subliminal manipulation techniques that could cause physical or mental harm, as well as the exploitation of vulnerable groups.
For high-risk applications, clear requirements and specific obligations are established for those responsible for deploying and providing the systems.
Additionally, the regulation highlights important points, such as:
- the requirement for a compliance assessment before certain AI systems are made available;
- the definition of requirements for governance, risk management, ethics, and transparency; and
- the imposition of sanctions for violations caused by AI systems.
The implementation of the regulations will occur gradually. It is crucial that companies seek guidance to comply with the requirements of the regulation. Even if they do not work with technologies that impact the European market, they should already take measures to ensure governance of AI initiatives in accordance with global standards.
For any questions regarding the topic, please contact TozziniFreire’s team.
