EU Council adopts pioneering AI act to ensure ethical development
12:02 PM EDT, May 23, 2024
The Council of the European Union has approved the AI Act, a law regulating matters related to artificial intelligence. These pioneering global regulations aim to ensure AI technology's safe and ethical development.
Why was the AI Act created?
As early as 2021, the necessity of introducing new regulations in the field of artificial intelligence was signaled. After long negotiations and resulting compromises, the introduction of the AI Act became inevitable. In March, the European Parliament approved the AI Act, and the Council of the European Union has officially adopted it.
What is the AI Act?
The AI Act is the first global act and may become a model for other countries and regions, similar to the General Data Protection Regulation (GDPR). This law stipulates that AI technology must comply with European Union values, protect fundamental rights, be human-centered, and serve as a tool aiding humans. It introduces a range of prohibitions and restrictions.
Key provisions of the AI Act
The AI Act introduces stringent obligations related to risk assessment. It classifies AI technologies into four categories: low, medium, high, and unacceptable risk. Technologies classified as unacceptable risk will be banned in the European Union.
The law mandates ensuring the transparency of AI systems, informing users about their capabilities and limitations, and protecting the rights of individuals affected by them. AI system operators must also have the appropriate competencies regarding their use and understanding of their operation's results.
The act also prohibits using AI systems to manipulate human behavior if it can lead to serious harm. It forbids using biometric systems to determine information about political opinions, union membership, religious or philosophical beliefs, race, sexual life, and sexual orientation.
Commencement of the AI Act
The obligations regarding general AI models will become effective 12 months after the law is introduced, while the provisions concerning AI systems embedded in regulated products will become effective after 36 months. The ban on biometric categorization will become effective 6 months after the new regulations are introduced.
In an interview with Reuters, Patrick van Eecke from Cooley law firm emphasized that the law will have a global reach. Companies outside the European Union that use EU customer data on their AI platforms will have to comply with this law. He added that other countries and regions will likely use the AI Act as a model, similar to the GDPR.
