U bent hier:

The world’s first AI Act: What can we expect?

As modern-day students, it won’t come as a surprise to you that Artificial Intelligence (AI) has become an integral part of everyday life. Generative AI services, such as ChatGPT and Gemini, are a frequent occurrence in- and outside of educational institutions. Maybe your teachers have warned you about the risks of using AI or perhaps they allow the use of it. I even could have asked ChatGPT to write this article for me if I wanted to. AI, and the rapid rate with which it is developing, is both revolutionary and concerning at the same time. Apparently, this concern has reached the European Union as well. Last month, the EU passed its first law to regulate the use - and misuse - of AI. 

The first AI Law 

On March 13th, the ‘Artificial Intelligence Act’ was approved by the European Parliament, with a vast majority of 523 out of 705 members voting in favour of the new act. This EU act is the world’s first binding law on artificial intelligence and regulates the use of different types of AI-powered systems throughout all 27 Member States. The AI act was first proposed in December 2021 by the European Committee, which is the EU organ that is responsible for preparing and proposing new laws to the Parliament. The reason why so much time has passed between the proposal and the approval of the AI act is because the EU wanted to include the relatively new ‘general purpose AI’, which is an umbrella term for programs such as ChatGPT, Gemini and Midjourney. However, these generative AI functions are not considered to be the most dangerous, and therefore will not be dealt with first. For now, the priority of the AI act is mostly to tackle technologies that track and manipulate civilians. A brand-new ‘European AI Office’ will provide guidance for governments and companies with the implementation of the new rules. 

European Dilemma 

The way to the final approval of the Artificial Intelligence Act was not an easy one. Before the start of the negotiations, experts from Brussels had estimated the odds of a successful negotiation as a 50/50 chance. However, after two days of intensive assemblies, the European Parliament managed to reach a decision and passed the new law. The hesitation is not misplaced, as creating the first AI legislation poses a huge dilemma for the EU. On one hand, the use and development of AI services should stay under the control of authorities and should thus be strictly regulated. On the other hand, the EU does not want to create a legislation that is too strict, as this will discourage European AI-developers from advancing their technology. A stagnation in European AI development would mean that Europe becomes dependent on the US and China for their AI services once again.


As significant as the legal developments around AI seem, the work is not done yet. First, the EU Member States have to vote on this law as well. If all goes well, this will happen in April. Only once Member States have passed the law and adopted it into domestic legislation, it will form an effective legal framework. Aside from this, a new European Committee will take office in a few months. The current Committee has stated that their successors will have to start preparing specific rules for different forms of AI, including ChatGPT and Gemini, right away. Until then, your AI generated homework is probably safe. 

Bronnen :

Europees Parlement stemt definitief in met regels voor AI | Trouw

EU legt kunstmatige intelligentie aan banden: hoe gevaarlijker de toepassing, hoe strenger de regels | de Volkskrant

EU Passes AI Act, World's First Comprehensive AI Law, By Overwhelming Majority | Legaltech News

From DMA to AI Act: Understanding the EU's Tech Legislation Landscape | Law.com International