European Artificial Intelligence Act: the state of the art

The European Commission's Initiative

Digitalization is playing an increasingly central role in the European economy. As a result, European institutions, starting with the European Commission, have embarked on a path to define clear regulations in the fields of technological innovation. 

In this context, the European Commission launched a significant initiative in April 2021 by publishing a draft proposal for a European regulation on artificial intelligence. Once approved, these rules would constitute the world's first legislation dedicated to artificial intelligence. 

Goals of the regulation

European institutions and Member States had long formalized the need to regulate the marketing and use of artificial intelligence systems. 

In particular, the European Commission identified a series of objectives and needs in its communication to the European Parliament, the Council, the European Economic and Social Committee, and the Committee of the Regions on April 25, 2018 (the "2018 Communication"). Among these, it emphasized the importance of ensuring that Europe is competitive globally in the field of artificial intelligence, that the use of artificial intelligence is inclusive, and that this new technology respects the fundamental values of the European Union.

Consequently, the main purpose of the regulation on artificial intelligence is to improve the functioning of the European internal market by establishing a uniform legal framework, particularly concerning the development, marketing, and use of artificial intelligence in accordance with the Union's values.

Regulatory approach

In determining the new regulations, European institutions follow a risk-based approach, establishing obligations for providers and users based on the level of risk that artificial intelligence can generate. In particular, facing a technologically neutral definition of artificial intelligence, different risk levels are identified: unacceptable risk, high risk, limited risk, and minimal risk. While artificial intelligence systems at unacceptable risk are prohibited (as a general rule), different obligations are associated with each other risk level. 

Following the recent proliferation of so-called Generative AI systems, capable of autonomously generating complex texts (e.g., Chat GPT), images (e.g., Midjourney), audio, and/or video, the European Parliament introduced obligations related to generative AI systems in drafting its proposed amendments to the Commission's draft regulation.

It is important to note that the newly produced regulations will not be limited to application to European companies and/or consumers but will also cover the supply and/or use of artificial intelligence systems within the territory of the European Union, regardless of the origin of the manufacturer or distributor. This element is aimed at ensuring that European companies do not suffer a competitive disadvantage but that companies from third countries, if they are interested in the European market, also respect its fundamental principles.

What to expect in the coming months

Following the European Commission's proposal, trialogues with the European Parliament and the Council were initiated in June 2023. 

On December 8, 2023, representatives of the European Parliament and the Council reached a provisional agreement on the draft European regulation on artificial intelligence. The text will have to be voted on and definitively approved by both institutions, which is expected to occur in the coming months.