Political speeches at international summits often trade in aspirations rather than specifics. Emmanuel Macron’s address at the AI Impact Summit in Delhi was unusual in that it contained both — and the specifics were concrete enough to be translated into policy. Across his remarks, Macron outlined a vision for AI governance centred on children that has real operational content: not just a direction but a destination.
The first element of his vision is the principle that the laws governing the physical world must extend to the digital one. If something is illegal in real life — sexually explicit content involving children, for instance — it must be illegal online. This sounds obvious but has rarely been enforced at scale. The Unicef-Interpol research that found 1.2 million children victimised by AI deepfakes in a single year is the most recent evidence of the gap between this principle and practice.
The second element is platform accountability. Macron called for platforms and governments to work together, framing this as a shared responsibility rather than an adversarial relationship. His argument is that platforms that profit from children’s engagement must accept legal responsibility for the safety of that environment. Voluntary commitments, the record shows, are insufficient. Legal accountability is the mechanism that makes responsibility real.
The third element is domestic legislation as a model for international coordination. France’s proposed ban on social media for under-15s is not presented as the final word but as a starting point — evidence that governments can act and a template for what action might look like in other national contexts. Through the G7 presidency, Macron wants to use this domestic precedent to drive international standards.
Fourth is international institutional architecture. Without naming a specific new body, Macron called for governments, platforms and regulators to develop coordinated governance mechanisms — a position that aligns with Sam Altman’s more explicit call for an international AI oversight body. And fifth, overarching all of this, is the political commitment of democratic governments to treat child safety as a non-negotiable priority rather than a consideration to be balanced against commercial interests. These five elements together constitute a governance vision that is more concrete than most of what the summit produced.
