European Commission: an AI regulation on fire

While the proposed regulatory text on artificial intelligence is at the heart of discussions between the European Parliament and the Council of the EU, Thierry Breton – European Commissioner for the Internal Market since 2019 – specifies the elements of this text. Based on four key points, this regulation should “take due account of the protection of security and fundamental rights, as well as the economic and societal benefits of technology”.

This regulation will be presented as the first certification system for AI with a legal framework and security for the citizens who use it through European or international companies. This will allow “European companies to be competitive on a global scale thanks to a marking system certifying a“ trustworthy AI ”,” explains Thierry Breton. The rules will apply uniformly across the 27 member states, and surprisingly, “will also have to apply to producers outside the EU. Because we have to make sure that the AI ​​systems placed and used on the European Union market respect our fundamental rights and comply with our values ​​and our rules ”. Through these rules, the European Commissioner is targeting Facebook, and more broadly, all the GAFAMs with a monopoly on data and its processing in the world.

Europe, a future key player in AI

The proposal also presents “a balanced and proportionate horizontal regulatory approach to AI that is limited to the minimum requirements necessary to address the risks and problems associated with AI”, so as not to restrict or hinder technological development or increase disproportionately the costs of bringing AI solutions to market. These obligations will be imposed on providers of high-risk AI systems. Proportionate obligations are also imposed on users and other participants located along the AI ​​value chain (e.g. importers, distributors and agents).

Obviously, with this proposal, the objective is to make Europe a key player in the development of secure, reliable and human-centered artificial intelligence systems. The fact remains that today the market for AI frameworks is dominated by American and Chinese players. In parallel, the Commission is working on the development of a Data Act, a regulation concerning access to data and aimed at improving “fairness in the data economy”. To this end, the deployment of common European data spaces is planned “in strategic sectors and areas of public interest, such as agriculture, health or mobility”.

DSA and DMA, two complementary texts

On the occasion of his intervention at Big Data 2021, Thierry Breton also mentioned the progress of two projects, the Digital Services Act (DSA) and the Digital Market Act (DMA) to organize the obligations of the major platforms involved in the European Union. The DSA should impose obligations on social networks such as the fight against illegal content, hate speech or disinformation. “There is the territorial space, the maritime space and now there is the information space. The latter is perhaps the space in which our children may even spend more time than in the first and yet there are not enough rules, ”believes Thierry Breton.

With this text, he assures us that “everything that is prohibited in physical space will be prohibited in information space and everything that is authorized in physical space must also be”. The platforms will also have the obligation to have a seat in Europe, whatever the type of establishment, in order to put an end to “European havens”, he adds. “Some large online platforms behave like ‘gatekeepers’ in digital markets. Digital market legislation aims to ensure that these platforms behave fairly online. The European Commissioner assures us that these three projects must be in line with the values ​​of the European Union. “It must correspond to our rules of rights and our values ​​and our rules of law are often based on our values” he concludes. This European regulation should be expected within 18 months.