The European Union (EU) Council declared on Tuesday that the artificial intelligence (AI) act of the bloc has received its final approval from ministers representing EU member states. Following the endorsement of the AI legislation by EU lawmakers two months ago, with significant amendments, the ministers have now cast their vote.
Belgian Secretary of State for Digitization Mathieu Michel stated that this law tackles a global technological challenge while also presenting opportunities for our societies and economies. As per the statement, the AI Act has implemented a risk-based approach, imposing more stringent regulations on AI applications with higher risks.
The legislation prohibits AI systems from engaging in cognitive behavioral manipulation and social scoring in the EU. It also forbids the use of AI for predictive policing based on profiling and systems that utilize biometric data to classify individuals by categories like race, religion, or sexual orientation. For general AI systems, compliance with transparency requirements is mandatory unless they present systemic risks. In such cases, stricter regulations must be followed. It is important to note that the EU’s AI Act makes exceptions for systems used solely for military, defense, and research purposes, exempting them from these rules.
To ensure effective implementation, the legislation establishes various governing bodies, including an AI office, a scientific panel of experts, an AI board with member state representatives, and an advisory forum for stakeholders.
Once signed by the presidents of the European Parliament and the Council, the legislative act will be published in the EU’s Official Journal and will enter into force 20 days after publication.