The main European rules to govern artificial intelligence on the continent are currently under discussion.
The European Union wishes, with the establishment of an AI Act, to once again build flagship regulation, this time in the field of artificial intelligence. Unfortunately for Brussels, this requires agreement among the various partners. This is why the meetings follow one another, like the one held this week. And if progress begins to be noted, it is time for the Brussels authorities to quickly find a definitive text to vote on!
Trilogues are multiplying
A new trilogue, a discussion between the European Commission, the Council of the European Union and representatives of the European Parliament, was held this week. It was the fourth this year to be organized as part of the future regulation of AI in our countries.
The co-rapporteurs of the AI Act Dragoș Tudorache and Brando Benifei as well as Thierry Breton hope to be able to find an agreement quickly, while a final trilogue is scheduled for the month of December. Because if at the end of the year a compromise was not found, discussions would have to continue at the beginning of 2024, which would put the bill at risk. The latter could in fact then have to go before a new legislature, installed after the European elections next June.
An agreement on Article 6
However, progress was made this week. Thus, the partners were finally able to find an arrangement on most parts of Article 6, the section which deals with artificial intelligence systems considered to be “high risk”. The European Union wants to establish several levels of dangerousness for AI, with the riskiest ones receiving the most sustained attention.
The parties involved did not want to give details of what had been adopted. “ We have made significant progress » simply indicated Brando Benifei. “ If the Council shows a constructive approach, we could approve (at trilogue level) by the end of the year » he hopes now.
Source : Reuters, Develop.com
0