Altman tells the G20 no child will ever outsmart AI: prophecy and lobbying in the same sentence

By Infobae · September 6, 2026.

Sam Altman, CEO of OpenAI, told the G20 Ministerial Meeting on Innovation, held in North Carolina, that "a child growing up today will never be smarter than AI." He said it at a forum of government officials, with US Secretary of Commerce Howard Lutnick among those present, as reported by Infobae. This is not an offhand remark in an interview: it is a message calibrated for those who decide regulation and public investment in technology.

To reinforce it, Altman turned to an analogy already classic in the rhetoric of AI labs: he compared rejecting artificial intelligence today with having rejected electricity more than a century ago. "I think it would be roughly as bad an idea... as it was to say we weren't going to have electricity," he said, according to the same source. He rounded out the argument with a ten-year prediction: society will stop "talking so much" about AI because it will be everywhere, as invisible as the electric current in a wall socket.

The rhetorical move is powerful, but it is worth viewing with the same distance we apply to any statement from an executive with a direct stake in the outcome. Altman is not a neutral observer: he runs the company that stands to benefit most from governments treating AI as inevitable infrastructure rather than as a technology that deserves regulatory friction. The comparison with electricity serves a double function: it normalizes adoption and, along the way, defuses the question of who controls and who benefits from that infrastructure, something that with the power grid did end up being regulated as a public utility.

The line about the child who "will never be smarter than AI" is, moreover, a prediction that is practically impossible to verify and that conflates different things. Intelligence is not a single scale on which a language model and a person compete for the same slot: it includes judgment, situated discernment, relationships, applied creativity and physical dexterity, areas where benchmark performance says very little. We have already seen at this publication how often a claim of "matching" or "surpassing" rests on a generous metric rather than on a broadly demonstrated capability.

That said, the real underlying problem Altman points to — what schools do when the machine knows more facts than any student — does deserve to be taken seriously, and that is where we believe the focus should be rather than on the headline. In our coverage of AI's impact on education we already argued that the winner is the teacher who orchestrates AI as a learning tool, not the one who competes with it by transmitting facts; that change of role, and not an intelligence race no child can win, is the hard, short-term transition education systems have to start managing now.

In the long run, Altman's reading connects with something we share: a truly ubiquitous and reliable AI, of the kind he describes for ten years from now, is compatible with the horizon of abundance we defend — more distributed capacity, less friction, resources and knowledge accessible to everyone. But that horizon is not guaranteed simply by sounding the alarm about children who will "never" catch up with machines; it is built with governance, a real sharing of the benefits and education that teaches people to use the tool, not on the premise that debating it is already a battle lost in advance.