The negotiated compromise on the European Union’s AI Act was unanimously approved on Friday.
Source: Photo by Steve Johnson on Unsplash
The EU Commission, the European Parliament and the member states have been negotiating the regulation of artificial intelligence in Europe for two years, and it is now clear: the compromise was unanimously approved in the Committee of Permanent Representatives of the Member States.
France and Italy outvoted
The federal government announced at the beginning of the week that after initial concerns it had agreed to approve the AI Act. Italy also ultimately agreed to the AI Act, but made their concerns known.
Only France continued to oppose the regulation in its current form, but eventually had to give in because they did not have the necessary blocking minority. This means that at least four EU states, which together represent at least 35 percent of EU citizens, are necessary to block a decision by the majority in the Council. In France, providers such as MistralAI and Poolside are supported by the government, which has led to widespread concerns about the regulation.
However, before the ordinance can finally be approved by the plenary session, it still has to go through a few stages. The next point of contact is the ministerial level in the Council of Member States. However, after the vote in the Coreper Committee, this is only considered a formal confirmation. The Civil Liberties (LIBE) and Internal Market (IMCO) committees in Parliament also still have to approve the result.
Largely clear path for the AI Act
However, it will certainly be some time before the law comes into full force. The transition periods set out in the AI Act will only begin when the law is promulgated in the Official Journal of the European Union. However, the members of the European Parliament assume a clear majority for the AI Act. After extensive concessions from the European Parliament to its member states, the way for the regulation is now likely to be paved.