Emmanuel Macron spoke on December 11 on the recent European law on artificial intelligence (AI Act). He fears that this regulation will put the Old Continent in even greater difficulty in the face of giants like China and the United States.
At the highest level of the State too, there is irritation over the young European legislation on artificial intelligence. Speaking in Toulouse on December 11 during a progress update on the France 2030 plan, Emmanuel Macron judged that Europe was doing itself a disservice by legislating too early on AI. “ I think this is not a good idea “, he said.
A text unfavorable to the interests of Europe?
The fear of the President of the Republic? Let the Old Continent put obstacles in the way that will end up putting it at a disadvantage in international competition – in particular against the United States and China, two countries in which the means, progress and ambitions are the highest. “ We are all very far from the Chinese and the Americans », Warned the head of state.
Emmanuel Macron’s remark comes as France sees some champions emerging in the field of artificial intelligence. First, Mistral AI, which recently completed a major fundraising round and published a second language model. Then Kyutai, a research laboratory driven among others by Xavier Niel, the founder of Iliad (Free).
More generally, France has a growing ecosystem and several names are standing out. LightOn (design of large language models), Aleia (platform providing AI software), Giskard (platform for evaluating the quality of AI algorithms), but also Dust (use of large language models) or Hugging Face (learning automatic).
“ We can decide to regulate much faster and much stronger than our major competitors, but we will regulate things that we will no longer produce or invent. “, he warned. A ” strange week “, according to him, that the one which saw Mistral AI innovate and raise funds, a sign of ” French genius as we like to see and celebrate it », and Brussels set legal limits.
A text on AI with 4 levels of risk
The AI Act saw significant progress on December 8 with a trilogue agreement between the European Commission, the Council and the Parliament. The text must, however, still be formally validated by the Parliament and the Council. Once this step has been completed, it will be directly applicable in the Member States. These steps can still take months.
Broadly speaking, the text intends to control the development of AI using an approach based on the risk of this or that system. Beyond generative AI embodied by services like ChatGPT or Midjourney, it also involves dealing with very sensitive issues: facial recognition, predictive policing, social rating, algorithmic justice, etc.
Four levels of risk for AI have been implemented: minimal, limited, high and unacceptable. Each level must follow specific rules, increasingly restrictive as one goes up the ladder of danger – the last level is supposed to lead to an outright ban on the tool at European level.
The text approved in trilogue is, however, contested — for the reasons mentioned by the Head of State, but also because the legislation would miss the most critical targets, such as facial recognition. This biometric process uses a person’s facial features, stored in a database, on which rules can be applied.
Evaluate the text and modify it if necessary
Emmanuel Macron, however, assures that he is not hostile to the idea of regulation in principle. “ This is a good thing » because she will “ consolidate a French model of regulation “, he conceded. However, the president pleads for regular evaluation. And, understand implicitly, for a quick correction if it turns out that the text imposes too many obstacles.
“France is undoubtedly the leading country in terms of AI in continental Europe”
Emmanuel Macron
“ France is undoubtedly the leading country in terms of artificial intelligence in continental Europe, we are neck and neck with the British. But they will not have this regulation on the foundational models », Underlined the Head of State. Neither are China and the United States, even if the lines are moving in Washington.
“ We must therefore always be at the right speed and in any case at the right rhythm. » in this “ongoing revolution”. Otherwise, we risk losing leaders and pioneers, he warned, and leaving other nations to take the lead in AI. In this war of AI there are also issues of power, governance, influence and model – including society.
