On Friday 2 February, the ambassadors of the Member States of the European Union (Coreper) will be invited to give their views on the agreement reached by the representatives of the EU Council at the beginning of December on the future Artificial Intelligence (AI) Act (see EUROPE 13311/17). In the meantime, certain countries, led by France, are still trying to influence certain aspects of the text in their favour.
Just under a week ago, the Belgian Presidency of the EU Council presented the final text to the Member States. This included the elements as agreed by the European co-legislators. On this basis, several national delegations asked for clarification on the issue of open source, exemptions and entry in to force.
With regard to definitions, several Member States stressed that ‘inconsistencies’ could lead to legal uncertainty. Beyond definitions, Paris and Berlin are said to have also wanted to return to the question of general-purpose AI systems.
In addition, the issue of copyright is important to France. On this point, the text could provide that the remuneration of players in the cultural sector, whose data is used to drive AI models, could exist, but that it would not, however, be made entirely public.
While the European Commission announced, on Wednesday 24 January, that it had adopted the decision to establish the future ‘European Office for Artificial Intelligence’ (see EUROPE 13335/6), in accordance with the provisions set out in the text, some Member States said that the mandate of the future body should be readjusted and scaled down. (Original version in French by Thomas Mangin)