February 10, 2025, Emmanuel Macron triumphantly announced an investment of 109 billion euros in “Action Summit on Artificial Intelligence” which he hosted in Paris. According to him, it’s equal “for France what the United States announced with Stargate”, a $500 billion plan launched by Donald Trump around OpenAI, the company that created ChatGPT. The President of the Republic has vowed to make France, which is experiencing deindustrialization, the number 1 host country in Europe for supercomputers and data centers vital for generative AI. Nine months later, it turns out that only investment projects worth 26 billion euros have been confirmed (including a super campus project dedicated to AI from the Emirates fund MGX), although Elysée is talking about 90 billion euros. “safe”. And the Court of Auditors gave public authorities both good and bad points in its report “national strategy for artificial intelligence” delivered this Wednesday, November 19 by its president, Pierre Moscovici.
It is rather the overall satisfaction that emerges from this 104-page document in which the “wise men” of rue Cambon estimate that “The national strategy for artificial intelligence has created real momentum since its launch in 2018.” And emphasize it “France has successfully positioned itself as a leader in Europe, including in the latest generative AI technologies.” A direct allusion to French champion Mistral AI which developed the Le Chat chatbot and just raised 1.7 billion euros. The epitome of French genius in the eyes of captivated macronies, the start-up led by Arthur Mensch is still far from being able to compete with the likes of ChatGPT (OpenAI) and Gemini (Google). “It can be compared with its competitors, apart from the United States and China,” politely noted the Court of Auditors.
But a 12 billion euro “unicorn” is taking advantage of ours “Your Honour” in “research and higher education in artificial intelligence”. While French “deep learning” pioneers like Yann Le Cun (Meta) or Yoshua Bengio (University of Montreal) left their country to train their talents across the Atlantic, France is currently slightly more successful in retaining its talent: “In terms of AI research and training, it ranks third in the world. More than 4,000 French researchers are working on AI at the moment.welcomes the Court. And in one year, “Our country rose from thirteenth to fifth position in the Global AI Index».
After Mistral, the number of AI startups in France has also doubled since 2021: “More than 1,000 of them will be active in this field by 2025 and they have raised almost 2 billion euros in funding by 2024,” note down the report. And “several large French groups are increasing their offerings and investments in AI research.” Cocorico, France is also the same “The leading European country in terms of the number of foreign investment projects in the field of artificial intelligence, and hosts Europe’s leading research and decision-making center for world leaders in the field of AI.”
But behind the beautiful appearance of the AI Summit, the judges at the top financial court certainly noted failure: by 2022, “Some of the announced priorities, particularly regarding the important issues of training or supporting change in the economic sectors most affected by AI, have not been implemented or are being implemented in a very limited way, thereby risking leaving France disadvantageously left behind.” Due to the emergence of generative AI “not anticipated ahead of the ChatGPT revolution”. And public support as part of the “France 2030” plan.» still small: 1.3 billion euros between 2018 and 2022, only 1.1 billion euros for the period 2023-2025, the reporters noted.
Race results: “The priority of supporting business demand for artificial intelligence solutions has only benefited from very modest measures, and the expected acceleration and expansion of the deployment of artificial intelligence in the economy has not yet occurred.” And we read further: “Important projects regarding schools and universities are still not designed and implemented”, “the transformation of public action through artificial intelligence also remains very disappointing”… Without speaking “actions aimed at a wide area and audience” which is still embryonic.
However, the Court of Auditors wants to be more optimistic by formulating ten recommendations “whose implementation will allow public policy regarding artificial intelligence to change scale” And from “exceeds identified limits and deficiencies”. At the top of the list, he advocates creationism “ad hoc secretariat general” to increase “interministerial management of AI public policy”, which will make some Elon Musk admirers in the ranks of “French Tech” chuckle. But authorities also recommend use “public procurement levers” For “strengthening research transfer towards industrial development and support for the growth of AI companies”which will certainly be more attractive to entrepreneurs in this sector. However, considering the state’s financial condition, this is more of a promotional issue “a new form of partnership between the public and private sectors”.
The Court of Auditors also recommended “to support the acceleration and expansion of AI adoption by companies in the next five years”from “strengthening access to data for AI systems”And “to invest in the country’s storage capacity”. He also suggested “build realistic ambitions in the issue of electronic components for AI” to reduce dependence on foreign countries when it comes to obtaining critical GPU chips. But it’s hard to see how France can survive without Nvidia and the rest of AMD given the recent setbacks of its champions STMicroelectronics.
We don’t know what public authorities will include in their list of “what we have to do is, we have to do it“, and any government caught between budgetary and political traps will not be able to implement these new policies “national strategy for AI” in the next few months. And if this report is deemed necessary “to adapt educational methods and tools as well as the content of all higher education courses with AI”, on the contrary, it ignores the alarming social and civilizational problems caused by artificial intelligence. We’re not asking an accountant to become a philosopher, but why not keep the coming revolution in mind.
