EU Approves First Global Regulations for Artificial Intelligence
The Council of the European Union has approved the AI Act, establishing the world’s first comprehensive regulations for artificial intelligence, aimed at harmonizing rules and addressing varying AI risks.
Introduction of the AI Act
The AI Act, approved on 21 May 2024, represents a significant milestone in global AI regulation. It introduces a risk-based approach to categorizing AI systems, ensuring that higher-risk applications meet stringent requirements while promoting innovation with minimal barriers for low-risk applications. This legislation is expected to set a global precedent for AI governance.
Key Provisions and Governance
The AI Act establishes a comprehensive governance framework to enforce its provisions. This includes the creation of an AI Office within the European Commission, a scientific panel of experts, an AI Board with representatives from member states, and an advisory forum for stakeholders. Each body plays a crucial role in ensuring the consistent application and enforcement of the regulations across the EU.
Risk-Based Categorization
AI systems under the Act are categorized by risk levels. High-risk systems, such as those used in critical infrastructure, healthcare, and education, are subject to rigorous requirements, including fundamental rights impact assessments and mandatory registration in the EU database. These systems must also adhere to strict transparency and accountability standards to mitigate potential harms.
Prohibited Practices
The AI Act explicitly prohibits certain high-risk practices. This includes AI systems used for cognitive behavioral manipulation, social scoring similar to practices seen in China, and predictive policing based on profiling by race, religion, or sexual orientation. These prohibitions aim to prevent misuse of AI technologies and protect fundamental human rights.
Support for Innovation
To foster innovation, the AI Act introduces regulatory sandboxes. These controlled environments allow developers to test and validate new AI systems without the immediate pressure of full compliance, encouraging experimentation and development of cutting-edge technologies. This initiative is particularly beneficial for startups and SMEs, providing them with a competitive edge.
Penalties for Non-Compliance
The Act stipulates substantial penalties for non-compliance, calculated as a percentage of a company’s global annual turnover or a predetermined amount, whichever is higher. This ensures that violations are met with significant consequences, thereby promoting adherence to the regulations.
Implementation Timeline
Following its publication in the EU’s Official Journal, the AI Act will enter into force 20 days later. Member states have a two-year period to transpose the regulations into national law, with full enforcement starting in the spring of 2026. This phased approach allows for a smooth transition and adaptation period for all stakeholders involved.
Bronnen
- euraxess.ec.europa.eu
- www.bundesregierung.de
- www.tagesschau.de
- www.lto.de
- www.consilium.europa.eu
- www.br.de