At the end of a political marathon, Europe has reached an agreement on the regulation of artificial intelligence. Associated with Germany and Italy, France hoped to tip the scales in favor of its softer line with regard to the so-called foundational models.
The trio was unable to make its voice heard sufficiently and influence the decision of the 27. In his speech on the occasion of the two years of the France 2030 recovery plan, President Emmanuel Macron did not hide his disapproval of the with regard to the agreement reached by the trilogue.
French genius thwarted by the AI Act
The French leader perceives a contradiction between two major events of last week. The first is the record fundraising of Mistral AI (385 million euros). “It really is French genius as we like to see it and celebrate it,” he congratulates himself.
The second is therefore the European regulation, the AI Act. For the president, this “has come to consolidate, and this is a good thing, a French model of regulation.” However, he considers it too restrictive on foundation models.
“We are the first place in the world where on so-called foundational AI models, we will regulate much more than the others. I don’t think it’s a good idea and I say that honestly,” he says.
Emmanuel Macron is therefore already calling for an evaluation of the AI Act, which according to him could penalize Europe in the international competition taking place with the British, the Chinese and the Americans.
Europe alone to regulate foundation models
“So 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. This is never a good idea. We must therefore always be at the right speed and in any case at the right rhythm,” he reacts.
To preserve the French position on AI and not widen the gap with China and the United States, which preserve these models, the president “asks that we evaluate this regulation on a regular basis.”
Emmanuel Macron identifies a particular criterion justifying an adaptation of the AI Act: “if we lose leaders or pioneers because of that, we will have to come back. This is key.” And the president also defends the government’s record in the artificial intelligence sector.
“We put 1.5 billion on our AI strategy. We have labs doing fantastic work, research organizations, notably INRIA (…) We have AI that we even use on our educational project,” he quotes.
If we lose basic control, we’re screwed
Emmanuel Macron also highlights the 500 million euros invested in the Jean Zay supercomputer. The success of a public policy on AI is based on fundamentals, he argues. These are computing capabilities, open source and startups.
“But if we lose basic control, we’re screwed. The basic control is to have data centers here, a real European cloud and to have the capacity to control the value chains,” he warns.
Finally, the French president calls on the scientific community, electronics companies and the various players in the sector to develop “a proposed roadmap allowing the emergence of components and system architectures for massive data processing in artificial intelligence, with the objective of an energy consumption which is divided by 100 or 1,000 compared to the current standards of current processors and graphics cards.”
“We need to win the battle for sober AI. This is key and this is where we can be different. Redoubled our efforts, France 2030 must allow us to secure this chain,” he explains. .