Since humanity’s earliest stirrings, education has always been the privileged instrument through which civilizations shape their destiny. From Mesopotamian scribes to Silicon Valley engineers, each era has forged its elites according to the imperatives of its time. But never before have we witnessed such a radical transformation as the one now unfolding in Chinese classrooms.
I observe, with that fascination mixed with unease that great historical shifts inspire, a silent revolution redefining the very codes of knowledge transmission. Starting in September 2025, every Chinese child, from the age of six, will discover artificial intelligence not as one subject among others, but as the new language of their era.
When Huai Jinpeng, China’s Minister of Education, describes AI as the “golden key” to the education system, his words echo with the resonance of great imperial proclamations. This metaphor reveals the Promethean vision of a country determined to outpace the West in the race toward this cardinal technology. For behind this educational reform lies a truth few dare articulate: whoever controls AI tomorrow will hold the keys to global hegemony.
In Chengdu, Xiaoming, age 6, begins his first coding lesson with Meiling, an educational AI with a gentle voice who will accompany him throughout his schooling. He doesn’t yet understand what “neural network” means, but he already knows how to train a model to recognize insects in his garden.
Meanwhile, in Montpellier, Lucien, age 6, has received his first login credentials to access Pix Junior. He’s learning to click an icon to open a word processor.
A power strategy of striking ambition
This transformation stems from no improvisation. It follows the implacable logic of a plan whose coherence commands admiration, even from its most convinced detractors. China, with that millennial patience characteristic of its strategic culture, has understood that the real battle will not be fought in today’s laboratories but in tomorrow’s minds.
The program’s architecture reveals remarkable sophistication. In elementary school, children tame AI through play. In middle school, emphasis shifts toward concrete application. In high school, the creative dimension flourishes, transforming users into creators. This progression conceals a vaster ambition: transforming every Chinese citizen into an “AI digital native,” creating the first society massively competent in this technology.
Unlike traditional elitist approaches that concentrate expertise in a few centers of excellence, China bets on capillary diffusion of skills. This massive democratization constitutes the most formidable disruption weapon of our era. While the West still debates the modalities of integrating AI into its school curricula, the Middle Kingdom is already training entire cohorts of children who will grow up with this technology as a natural companion. In ten years, these young Chinese will master AI as we master computing, with that instinctive ease that early learning confers.
The figures already testify to this strategic advantage: China produces nearly half of the world’s top AI researchers, compared to only 18% for the United States. To support this transformation, Beijing deploys colossal resources: one hundred expert teachers and one thousand specialized teachers trained for the capital alone. This mobilization recalls the great imperial works, when the Middle Kingdom erected its walls. Today, it’s building a cognitive wall for conquest.
Xiaoming creates his first chatbot to help his grandparents manage their small business.
Lucien, meanwhile, discovers that copy-paste on a keyboard is written “Ctrl + C”.
This transformation unfolds within a geopolitical context where AI is progressively becoming an issue of national sovereignty. Goldman Sachs estimates reveal the scale of what’s at stake: AI could contribute 0.2 to 0.3 percentage points to Chinese GDP by 2030. These decimals represent astronomical sums that will redefine global power balances. In January 2025, Beijing organized an international seminar on education in the intelligent era, bringing together three hundred participants from around the world—a soft power exercise of rare subtlety whose message is crystal clear: we’re not content to develop our model, we’re exporting it.
The ambition exceeds mere technical training. The program aims to develop algorithmic thinking, collaborative skills, ethical sense, and that spirit of innovation that distinguishes creators from simple users. China isn’t just training technicians but a generation capable of thinking in AI and acting with AI.
… And meanwhile, in France, we have PiX
Saturday, June 14, 2025. After four months of intensive consultation that mobilized… 500 contributions. Five hundred. Out of the 1.2 million people in National Education. That’s 0.04% of the educational community. A sample of striking representativeness, composed essentially of overwhelmed teachers who found time between grading papers and worried parents still wondering about the difference between WiFi and artificial intelligence.
But finally, France has decided “once and for all” on AI use in schools. The verdict? Generative AI will be authorized… starting in eighth grade only. And under strict conditions, naturally. This announcement, made in the discretion of a summer Saturday, reveals the French art of transforming a technological revolution into administrative reform.
The French system unveils an architecture of thoroughly Cartesian prudence. In elementary school, students “discover AI mysteries without touching interfaces“—understand before using, a touching philosophy. In middle school, finally, authorization to “dialogue with algorithms,” but within a “strict pedagogical framework.” In high school, autonomy becomes “possible,” but only in projects “explicitly defined by teachers.”
The crown jewel of the system? Mandatory training for all eighth and tenth graders, lasting a… dizzying duration: between 30 minutes and 1.5 hours. MAXIMUM. The length of an extended recess to master “prompting basics” (translation: how to ask AI questions without looking ridiculous), understand how generative AIs work, and naturally, environmental impacts. Because we must remind young people that every ChatGPT question melts an iceberg somewhere.
This millimetric progression starkly contrasts with Chinese Promethean ambition. While little Chinese children spend twelve years growing up with AI as a natural companion, France bets on an hour and a half of personalized training to create experts in “ethical and sustainable prompting.” Where Beijing deploys its “one hundred expert teachers and one thousand specialized teachers” to train “AI digital natives,” Paris organizes consultations and proposes micro-modules.
The French framework, in its administrative wisdom, also bluntly reminds that “generative AI devours enormous amounts of energy and water” and encourages “digital sobriety.” A laudable ecological concern that apparently doesn’t trouble Beijing strategists’ minds, too busy shaping tomorrow’s technological hegemony. Because “sometimes, a classic web search amply suffices to solve a given problem“—a formula that aptly summarizes hexagonal ambition.
In 2035, Xiaoming might create DeepSeek’s successor, a language model more powerful than ChatGPT.
Lucien will undoubtedly know that AI consumes lots of water, that sources must always be cited, and that sometimes Google suffices. He’ll also have perfectly mastered “prompting basics” after his intensive hour-and-a-half training. And above all, he’ll never have cheated by using ChatGPT without telling his teacher.
This French approach to educational technology perfectly illustrates that truth Talleyrand might have formulated with his customary mischief: “France always arrives on time, but never early, preferably after four months of consultation and a regulatory framework in triplicate.“
In 1997, Claude Allègre, Minister of National Education, declared his intention to “trim the mammoth’s fat.” Today, in the age of artificial intelligence, perhaps it’s the mammoth’s intelligence that needs trimming…