There are many tests for AI governance. Whether it prevents market concentration. Whether it enables democratic accountability. Whether it protects workers from displacement. But Emmanuel Macron has identified the test he considers most fundamental: whether it protects children. In Delhi, at the AI Impact Summit, he made this case with the authority of a head of state, the evidence of Unicef-Interpol research and the political platform of France’s G7 presidency.
The evidence is disturbing. Over 1.2 million children in 11 countries had their images manipulated into sexually explicit deepfakes by AI systems in a single year. In some nations, one child in 25 had been affected — one child in every classroom. The technology that enables this abuse is legal, available without restriction and improving rapidly. Macron’s argument is that a governance system that cannot prevent this is failing at its most basic level, whatever else it might accomplish.
His response is both domestic and international. France’s proposed ban on social media for under-15s reflects a national commitment to protecting children from environments that are demonstrably harmful. The G7 presidency gives Macron the platform to make this commitment international, pushing for enforceable standards that would require platforms and AI developers to take legal responsibility for child safety. He rejected the notion that this is anti-innovation — arguing instead that innovation built on the exploitation of children is not innovation worth having.
The Delhi summit brought together a remarkable array of voices on AI governance. António Guterres warned of dangerous concentrations of AI power. Narendra Modi called for child-safe, open-source AI development. Sam Altman advocated for an international oversight body. Dario Amodei raised concerns about autonomous AI systems. Each of these interventions touched on something real, but Macron’s focus on child safety gave the debate a moral urgency that the others lacked.
The question that follows Macron from Delhi is whether political alignment on child safety can be converted into enforceable policy. History suggests that international commitments without enforcement mechanisms produce little change. Macron’s domestic record — the willingness to pursue legislation rather than relying on voluntary industry action — suggests he understands this. The children who need protection are the test. They will tell us, eventually, whether the governance system that emerged from summits like Delhi actually worked.