Six months after its appointment by Matignon, the generative artificial intelligence committee will submit on Wednesday to the President of the Republic, Emmanuel Macron, its recommendations for adapting France’s strategy on these technologies that are absolutely crucial for its future. Chaired by the economist Philippe Aghion and Anne Bouverot, the president of the board of directors of the ENS, this committee brought together fifteen personalities with varied profiles, to best understand the subject through its different prisms. Among them, globally recognized researchers such as Yann Le Cun (head of AI research at Meta), Luc Julia (scientific director of Renault), Joëlle Barral (director of AI research at Google Deepmind), start-up leaders French companies specializing in the development of large language models (like Arthur Mensch of Mistral AI) or large technology companies (like Bernard Charlès, president of Dassault Systèmes). After 600 hearings conducted since October and a public consultation with 7,000 citizens, the committee makes 25 more or less concrete recommendations covering six major areas around artificial intelligence: the problem of training, the issue of the computing power necessary to operate all these technologies, those of access data without which there is no AI, the development of public research, the global governance of AI and the issue of financing innovation.

On this last subject, the report estimates that France needs to triple the level of investments currently made in generative AI technologies, to reach at least five billion euros per year. For this, the committee recommends in particular moving forward on the subject through the mobilization of savings capacities and strengthening the attractiveness of large funds on French territory to better finance start-ups.