France has made a spectacular leap in the Global AI Index, moving from thirteenth to fifth place globally in one year, and even reaching the third spot on the podium for research and training.
However, the report published on November 19 by the Court of Accounts presents a harsh reality: behind these flattering figures, the national strategy for artificial intelligence remains incomplete and too timid given the scale of the challenge.
The government officially established a three-phase division as part of the “National Strategy for Artificial Intelligence” (SNIA), launched in 2018 and led by the Interministerial Digital Directorate (DINUM) and the General Secretariat for Investment (SGPI). This structure aims to organize the progressive rollout of AI; here is what the report says:
Phase 1 (2018-2022): Ambitions scaled back from the start
The Villani report of March 2018 had nevertheless mapped out an ambitious path.
Seven years later, barely 20% of the issues it identified were actually addressed during the first phase. Defense, national security, and major economic shifts were largely neglected.
Promises of massive training and support for workplace transformations remained wishful thinking, exposing France, according to the Court, to a delay that could have serious consequences.
Phase 2 (2023-2025): Economic rollout misses its target
The second phase, which was intended to trigger the widespread adoption of AI by businesses and transform public action, is considered a failure. Support mechanisms are judged to be too modest and too confidential.
Artificial intelligence remains the preserve of a small circle of specialists and large companies. SMEs, artisans, local territories, schools, and universities have been left out. The long-awaited mass effect never materialized.
Read also: Yiaho awarded OpenAI’s ‘Tokens of Appreciation’ award for 10 billion tokens processed
Phase 3 (since February 2025): A change of scale is now vital
A new stage began in early 2025, with the priority goal of finally accelerating adoption by businesses.
But the Court of Accounts strongly reiterates: the context has radically changed since 2018. AI is no longer a technology reserved for experts; it permeates all parts of society. This reality requires a change in scale and method that current approaches cannot achieve.
Blind spots weakening the French position
Several essential areas have been neglected: anticipation of labor market disruptions remains almost non-existent, investment in data—the essential fuel for AI—remains insufficient, the transformation of public action and education is stalling, and no clear, realistic, quantified ambition has yet been defined. Recommendations for a strategy that finally meets the challenge
The Court of Accounts makes ten specific recommendations: strengthen true interministerial leadership, better coordinate European and local initiatives, conduct an honest evaluation of previous phases, sustainably anchor the research-training ecosystem, change the paradigm by putting AI at the service of the common good, and put national sovereignty back at the heart of the system.
The final message: no more half-measures possible
The conclusion is clear. Only a policy of a completely different scale will allow France to maintain its trajectory of excellence and embrace all the social, economic, educational, and strategic dimensions that the artificial intelligence revolution demands.
In the service of the common good and national sovereignty, the time for half-measures is well and truly over.
Source:
