EU's AI Act gets green light
.jpg)
European Union member states have given the final go-ahead to the AI Act, the regulation that sets rules for the use of artificial intelligence. The act is intended to respond to a global technological challenge that at the same time creates opportunities for European societies and economies.
The new act had already received majority support from the European Parliament, but now that the Council has given formal approval, the legislative process is over. European lawmakers speak of historic rules that protect against the risks of AI while leaving room for innovation. They hope the law sets a global standard.
The act relies on a risk-based approach. This means that the higher the risk of harm to society, the stricter the rules are. Low-risk AI systems are subject to very light transparency obligations, while high-risk AI systems must comply with a range of requirements and obligations to gain access to the EU market. Cognitive behavioural manipulation, social scoring and other risks considered unacceptable are banned.
Similarly, the use of AI for predictive policing based on profiling and systems that use biometric information to infer the race, religion or sexual orientation of a particular person are also not allowed. Separate rules have been created for generative AI, which can create its own content, especially in terms of transparency. Systems used exclusively for military, defence or research purposes are outside the scope of the act.
“We have struck a balance between encouraging the use of this rapidly changing technology and respecting the fundamental rights of our citizens,”said Belgian secretary of state for Digitalisation Mathieu Michel.
Carrefour Belgium tests an autonomous delivery robot of DELIVERS.A, equipped with artificial intelligence © BELGA PHOTO ERIC LALMAND
Related news
