Emmanuel Macron is a politician who reads statistics carefully, and the statistics he brought to the AI Impact Summit in Delhi carried a clear warning: act now, because if current trends continue, the numbers will only get worse. Research by Unicef and Interpol found that 1.2 million children in 11 countries had been victimised by AI-generated explicit deepfakes in a single year. As AI tools become more capable, more accessible and more widely distributed, that number will grow unless governments intervene.
That urgency animated every element of Macron’s Delhi speech. His defence of European regulation was urgent — not the measured advocacy of a politician comfortable with the status quo but the argument of someone who believes the window for effective intervention is closing. His call for international coordination through the G7 was urgent — not a long-term aspiration but a near-term political commitment. His domestic policy — legislation to ban social media for under-15s — is urgent in the sense that it represents action already under way, not a plan for the future.
The urgency is justified by the technology’s trajectory. AI tools for generating synthetic imagery are improving rapidly. The barrier to their misuse is falling. The scale of their deployment is growing. In this environment, governance frameworks that are developed slowly and enforced laxly will always lag behind the capabilities of the technology they are meant to govern. Macron’s argument is that catching up requires governments to move faster than they are accustomed to moving — and that the cost of not doing so is paid by children.
The support Macron received in Delhi from Guterres, Modi and parts of the tech industry reflects a growing awareness that the urgency is real. António Guterres’ warning that no child should be a test subject for unregulated AI carries exactly the right sense of immediacy. Modi’s call for child-safe technology by design acknowledges that the question is not whether to address this but how quickly. Sam Altman’s endorsement of international oversight, however cautious, suggests that even within the industry the case for urgent action is becoming harder to dismiss.
Macron’s plea in Delhi is ultimately a simple one: the evidence is in, the tools are available, and the time to act is now. The children who were victimised by AI deepfakes last year were not protected by deliberation. They needed and deserved action. The children who will be victimised next year, unless governments move, are waiting for the same thing.