The European Parliament has reached a political agreement with the Council on the legislative package for artificial intelligence (AI), known as the "Artificial Intelligence Act." It is the goal is to enhance the safety of AI use, protect fundamental rights and democracy, while also fostering innovation. The regulations establish obligations for AI systems based on potential risks and impacts.
Key Points of the AI Act Agreement
Some of the crucial points of the agreement include:
- Prohibited Applications: There are prohibitions on biometric categorization systems with sensitive features, collecting facial images from the internet for facial recognition databases, emotion recognition in the workplace and educational institutions, social scoring, and manipulation of human behavior by AI.
- Restrictions on Law Enforcement: Security measures and exceptions are in place for the use of biometric identification systems by law enforcement. Specifically, limitations exist in the areas of "real-time" and "post-remote" identification.
- Obligations for High-Risk Systems: AI systems with significant potential for harm are subject to clear obligations. This includes a mandatory assessment of their impact on fundamental rights. This also applies to the insurance and banking sectors. Citizens have the right to file complaints about AI systems and receive explanations for decisions.
- Guidelines for General AI Systems: General AI systems must meet transparency requirements, including technical documentation and compliance with EU copyright law.
- Measures to Support Innovation and SMEs: The promotion of "Regulatory Sandboxes" and real-world testing is encouraged. This facilitates and supports entrepreneurs, especially small and medium-sized enterprises (SMEs), in advancing the development of AI solutions.
- Sanctions: Non-compliance with the rules may result in fines for companies. The fine amount can be up to 7% of global revenue or 35 million EUR. This depends on the severity of the violation and the size of the company.
The AI Act text is not final yet. The agreed-upon text must now be formally accepted by both the Parliament and the Council to become EU law. Parliament's Committees on the Internal Market and Civil Liberties will vote on the agreement in an upcoming session.
Having further questions on this topic?
If you have further questions about AI and law or want legal advice on similar issues related to artificial intelligence, feel free to explore our services or contact us via email at office@geuer.at or by phone at +43-1-4380072. We look forward to addressing your inquiries.