A wind of panic is blowing over the AI Act. Negotiations on the draft European regulation on artificial intelligence (AI) have been very tense since MEPs walked out of a meeting with member-state representatives on Friday, November 10, Euractiv has reported.
The reason for their wrath? The elected representatives had just learned that France and Germany, supported by Italy, were opposed to a compromise on "foundation models:" the three countries no longer wished to regulate these artificial intelligence systems which can create text or images, and which can be used by companies in applications such as the ChatGPT or Bard chatbots.
Disagreements are common at trilogue meetings (the negotiation of a final text between the European Parliament and the Council of Member States, mediated by the Commission). However, this stumbling block could jeopardize agreement on the text, which is subject to a tight deadline. Furthermore, France's attitude is attracting criticism.
According to our information, in an attempt to find a way out of the crisis, Paris, through its Minister Delegate for Digital Affairs Jean-Noël Barrot, will be suggesting a set of minimum transparency obligations for foundation models, such as the publication of a summary of the software's characteristics, how it works and how it was trained. These rules could be brought together in a code of conduct adopted by companies. This idea is due to be put to the other member states at the next Council meetings, on Friday, November 17 and Tuesday, November 21.
'EU risks squandering historic opportunity'
With this offensive, France, Germany and Italy are opposing the obligations laid down by the European Parliament in the spring. The European Parliament called for regular assessments and mitigation of the risks posed by these models to the security and fundamental rights of the European Union (EU), including the fight against errors, illegal content, and discriminatory bias. These countries consider these procedures too cumbersome and costly for European start-ups. They also rejected the compromise put forward at the beginning of November, which envisaged applying these obligations only to the most powerful models.
Paris claims not to be opposed to AI regulation and believes that the AI Act would already be the most restrictive text in the world. France argues that foundation models will be subject to strong obligations if they are used by companies for "high-risk" uses, which form the initial basis of the AI Act. The uses in question are those in areas such as autonomous driving, exam grading, CV sorting, health diagnosis, granting credit, and justice.
You have 50% of this article left to read. The rest is for subscribers only.