On September 22, standing before the United Nations General Assembly, President Donald Trump announced that the United States would stop calling it artificial intelligence in official documents. From now on, the government’s word for the technology is super intelligence, SI for short, and he invited the rest of the world to follow.
His reasoning was pure Trump, which is to say it was about branding: “The use of the word artificial makes intelligence fake. It makes it sound fake, and it is not fake. It’s actually amazing.”
“From this point forward, all of United States documents and hopefully the world’s will be changed to use the much more accurate term super as opposed to artificial,” he continued.
The announcement landed the way most of his announcements land: half the room nodded, the other half reached for a stronger word than skepticism. Commentators called it absurd and dystopian. Senator Bernie Sanders’ circle pointed out the obvious definitional problem. Nobody at OpenAI, Anthropic, or Google has suggested they will change a single letter of their own vocabulary.
And yet. Strip away the messenger, and the message deserves a hearing. Because “artificial intelligence” was always a bad name, and “super intelligence” is genuinely better. We intend to use it in these pages, and this is why.
Artificial was never the right word
Think about what “artificial” means everywhere else in your life. Artificial sweetener: the lesser copy. Artificial flowers: the ones that fool nobody up close. Artificial turf: fine until you want the real thing. In every other context, “artificial” means an imitation of the genuine article, and imitations are, by definition, not to be fully trusted.
Now consider what we are actually being asked to trust this technology with: our schedules, our money, our health questions, our businesses, our creative work. The gap between the word and the weight is enormous. “Artificial” quietly tells the public that the thing is fake, lesser, a trick. That single word has been taxing adoption for years, and it has been doing it invisibly, which is the worst kind of tax.
Trump’s phrasing was characteristically blunt, but the underlying observation is one that linguists and product designers have made for decades: names are infrastructure. The label on the thing shapes what people expect from the thing, and expectations shape behavior.
The honest objections
Before we make our case, the counterarguments deserve their due, because a publication with a point of view should still be honest about the other side.
First, the rename currently applies only to United States government documents, and even there the machinery is unclear: as of this writing, no White House implementation order has been published. It is an announcement, not yet an apparatus. The private sector, which coined and spread the term “artificial intelligence” over seventy years, shows no sign of abandoning it on the strength of a UN speech.
Second, and more seriously, “superintelligence” already means something. In AI research and policy, it denotes systems that exceed human intelligence across most domains. Senator Sanders’ new bill uses exactly that technical definition while proposing severe penalties for building such systems unsafely. Collapsing that precise term into a general label for all AI muddies language that safety researchers genuinely need. That is a real cost, and we should not pretend otherwise.
Third, the timing invites cynicism. Days before the announcement, Trump polled the public on alternatives like “Superior Intelligence” and “Supreme Intelligence,” then chose a winner that was not even on the ballot. Critics are entitled to ask whether this is about clarity or about putting a personal stamp on the defining technology of the century.
All fair. Now here is why we land where we land.
Super says what the technology actually is
What does this technology do for the people who use it well? It does not imitate their intelligence. It extends it. It multiplies their output, sharpens their research, drafts at their direction, works while they sleep. We have written in these pages about the shift from the chat era to the agent era: the real story of AI in 2026 is not conversation, it is delegation, handing real tasks to software that completes them. Nothing about that is artificial. It is the opposite of artificial: it is capability, added to yours.
“Super” captures that. A superpower is not a fake power. It is your power, amplified. The name tells the truth about the relationship between the human and the machine: the human remains the principal, and the intelligence at their disposal is greater than what they were born with. That is not marketing gloss. That is the most accurate short description of the technology’s role in a well-run life.
And accuracy matters because expectations compound. A person who believes they are dealing with a clever imitation uses it like a toy: a smarter search box, a party trick for drafting emails. A person who believes they are dealing with super intelligence, an extension of their own mind, learns to delegate to it, to aim it, to build systems around it. The second person accumulates leverage. The first accumulates chat logs. In a decade, the distance between those two people will be very wide, and it will have started, in some small part, with what they called the thing.
This is the cultural capital argument we keep making: understanding a shift early, while it is still a choice, is worth more than understanding it late, when it is a requirement. Names are part of that understanding. Call it fake long enough and people will treat it as fake, right up until the moment it is running their competitor’s company.
On the definitional objection
What about the researchers’ term? We take it seriously, and we still think the public meaning will win, because the public meaning almost always wins. Language belongs to its speakers, not its coiners. Besides, the trajectory points in one direction: systems that exceed human capability in more and more domains. A name that anticipates where the technology is going is not the worst sin in nomenclature.
The government-documents limitation is similarly temporary in spirit. Official language has a way of leaking into journalism, then into product marketing, then into everyday speech. Whether or not a formal order ever appears, the debate itself has done something useful: it forced everyone to notice that they had been calling the most important technology of their lifetime by a name that means “not real.”
So we are switching
Here is our position, stated plainly. The technology is not fake. It is not an imitation of intelligence. It is intelligence placed at human disposal in a quantity no previous generation enjoyed, and the honest name for that is super intelligence.
We did not arrive here because a president gave a speech. We arrived here because the old name was doing quiet damage to public understanding, and the new name tells the truth. From here on, when we mean the field in general, that is what we will call it.
Welcome to the era of super intelligence. It was never artificial. It was always going to be super.
