Super intelligence: why Trump wants to rename AI
Super intelligence is the new name Trump wants to give AI. Understand what he said at the UN, what the word really means and what this changes.
NinaKalkazu.org
On Tuesday, September 22, 2026, Donald Trump took the podium at the UN General Assembly in New York and announced that artificial intelligence would get a new name. From that point on, he said, U.S. government documents would call the technology super intelligence, or "SI". The reason: the word "artificial" would make intelligence "seem fake".
It sounds like just another soundbite, and in part it is. But the name change touches on two words that have history and technical meaning. And it comes in the middle of a serious fight over whether to slow down or speed up AI.
In this article you will understand what Trump actually said, where the name came from, what "artificial" and "superintelligence" really mean and why this matters to anyone using AI.
Quick summary: the answer in 30 seconds. Trump said at the UN that American documents will swap "artificial intelligence" for "super intelligence", because "artificial" would sound fake. It was a speech, not a law, and does not apply outside the U.S. government. But "superintelligence" is already the technical name for an AI far more capable than any human, which does not yet exist. Applied to today's AI, the word exaggerates what it is.
Index
- What Trump said about super intelligence at the UN
- Does "artificial" mean fake?
- What is superintelligence, really?
- Why changing the name matters for AI regulation
- What this changes for you
- Frequently asked questions
- A name bigger than the technology
What Trump said about super intelligence at the UN
Trump announced that, "from now on", all United States documents, "and I hope the world's", will use "super intelligence" instead of "artificial intelligence". He said the new term "sounds much better" and is "much more accurate". And he admitted the doubt: "Let's see if I have any power. Maybe I do, maybe I don't".
The justification came next: "The use of the word artificial makes intelligence seem fake. And it's not fake. It's actually incredible". The sentence is in the UN's own coverage, and The Verge transcribed the entire passage.
The statement came in the middle of a long speech, whose main theme was Iran. The section on AI began with a refusal: the United States rejects "any attempt to build a globalist control scheme" for the technology. In the same sentence, Trump made the renaming, and the White House published the passage with the expression "henceforth officially called 'Super Intelligence'".
Then came the sentences that summarize the government's view. "Whoever wins in super intelligence wins", Trump said, repeating a slogan he previously used with "AI". According to him, the United States leads "by a wide margin" over China, and he will not "stifle the growth" of something "bigger than the Industrial Revolution".
Where the name came from: the Truth Social poll
The name came from two polls Trump conducted on his own social network, Truth Social, in the previous days. And "super intelligence" was not even in the first one.
On September 19, Trump wrote that "many people think" the term "Artificial Intelligence" is inaccurate and inelegant. In the first poll, the options were Superior Intelligence, Extreme Intelligence and Supreme Intelligence. "Superior" won with 41% of 66,041 votes.
Two days later, on September 21, came a second poll, now with "Super Intelligence" against "Superior Intelligence". "Super" won with 65% of 50,707 votes. The next day, the name was at the UN podium.
Worth noting the detail: Trump's followers voted on a network created by him. There was no consultation with researchers, linguists or technical agencies, as TechCrunch noted when covering the first vote.
Does this count as law?
No. So far it was an announcement in a speech. No decree (the executive order, signed by the president) has been issued changing the name, and Wired observed that it was not clear if one would come.
Trump has changed names before, and the path is usually this decree. That is how it was with the Gulf of Mexico, which became "Gulf of America" by Executive Order 14172, from January 2025. In August 2026, another decree renamed Lake Ontario "Lake America".
Even with a decree, the change applies to the U.S. federal government: its documents, maps and communications. No one outside the United States, in companies or universities, is required to follow.
In one sentence: it was a renaming made in a speech, inspired by a social network poll, that for now can only apply, at most, to U.S. government papers.
Does "artificial" mean fake?
No. In "artificial intelligence", "artificial" means made by people, constructed, and not born. It is the same sense as "artificial satellite" or "artificial lake": the satellite exists and really orbits, only someone manufactured it. The term never meant that intelligence is fake.
The expression was born in 1956, at a summer meeting at Dartmouth University in the United States. The organizer, mathematician John McCarthy, wrote in the proposal that the group would start from the idea that "every aspect of learning or any other feature of intelligence" could be described so well that "a machine can be made to simulate it". Dartmouth keeps this history as the birth of the field.
Notice the word "simulate". The researchers knew they were building machines that imitate parts of human thought. "Artificial" marked the origin, not the quality.
Think of artificial sweetener. It really sweetens, and the coffee becomes sweet. The name only warns that it did not come from sugarcane, it came from a laboratory. No one says the coffee "seems falsely sweet". With AI it is the same: the text that ChatGPT writes exists, the answer is there, and "artificial" only tells where it came from.
Trump is right on one point: words change perception.
For a long time researchers have discussed whether "intelligence" is the right term for programs that predict the next word in a text, like the phone's autocorrect that suggests the next word, only on a giant scale.
The criticism from these researchers goes in the opposite direction from Trump's: for them, the name already promises too much. Trump wants it to promise even more. The article about whether artificial intelligence can think like humans shows what these systems do and do not do.
In one sentence: "artificial" says that intelligence was manufactured, not that it is fake.
What is superintelligence, really?
Superintelligence is the name of a hypothetical AI, far more capable than the brightest people at practically everything, from science to strategy and social relations. It does not exist today. The term was defined and popularized by philosopher Nick Bostrom and became a centerpiece in the debate about AI risks.
The definition Bostrom published in 1997, in an article still on his website, says: an intellect "much smarter than the best human brains in practically every field, including scientific creativity, general wisdom and social skills". In 2014 he turned the idea into the book Superintelligence: Paths, Dangers, Strategies.
The detail that matters is "practically every field". A calculator does math better than you and is not superintelligent for that. Superintelligence would be beating any human at any mental task, at the same time.
That is why Bostrom's book became famous for risks, not promises. His question was: if we create something smarter than us at everything, how do we ensure it follows our goals? This discussion is where warnings from researchers and executives that AI could get out of control come from. In the UN speech, Trump put those making these warnings in the same group as the "hoaxes" he attributes to the opposition.
Superintelligence and AGI: what is the difference?
In industry vocabulary there are three steps. The AI you use today is good at specific tasks. AGI (artificial general intelligence) would be an AI as capable as a human being at almost everything. Superintelligence comes after: it would go far ahead of us.
| Today's AI | AGI | Superintelligence | |
|---|---|---|---|
| What it does | Specific tasks: writing, translating, programming, driving | Almost everything a person does, at a person's level | Everything a person does, much better than any person |
| Does it exist? | Yes | Disputed: part of the industry says it has arrived, part says it has not | No |
| Example | ChatGPT, Claude, Gemini | An assistant that learns any profession on its own | Something without a real example, by definition |
The boundary between steps is fuzzy, and the industry fights over it.
On September 14, in Los Angeles, the hosts of the American podcast All-In held their annual technology conference, the All-In Summit. On stage, Jensen Huang, CEO of Nvidia, said he thinks the industry has already reached AGI. When one of the hosts asked if we are already at superintelligence, Huang answered yes in "narrow segments", like a self-driving car that drives better than a human. The official podcast episode got the title "Superintelligence is Here".
Notice that this contradicts the original definition: "narrow" superintelligence is precisely what Bostrom left out. A car that drives well does not write poetry or negotiate a treaty. It is the calculator again, with wheels.
Why companies love the word
"Superintelligence" became the buzzword in the industry, and not by accident. It promises more than "AI" and more than "AGI", and that helps attract investors, talent and attention.
Meta, owner of Facebook and Instagram, named its research area Meta Superintelligence Labs, which has already released AI programs that create images and videos, according to the company's research page. There is even a company called Safe Superintelligence, which describes itself as a laboratory with "one goal and one product: a safe superintelligence".
In an episode on the topic, the Wired team called much of this marketing: vague and elastic terms, used to make technology sound impressive, valuable and intimidating, to the public and to lawmakers. The swap proposed by Trump enters this same current, now with a government's stamp.
In one sentence: superintelligence is the name of an AI that does not yet exist and that the industry increasingly uses as a brand for the AI that already exists.
Why changing the name matters for AI regulation
Because names change the perception of risk. Calling today's systems "super intelligence" makes them seem more ready, more powerful and more decisive in the competition between countries. At the same time, the government making the change rejects international rules and calls warnings about the technology a hoax. The new name reinforces the message: this is too big to slow down.
The contradiction becomes clear when you put the two ends together. In Bostrom's tradition, researchers talk about superintelligence to ask for caution. Trump uses the same word to ask for the opposite: "Let's stimulate, not contain", he said in the speech. As for risks, the promise was to watch the industry "closely" through the Department of Justice (the American ministry that prosecutes those who break existing laws), without new law.
You can see the effect in practice: both at the UN and at the All-In Summit, the grand name appeared attached to the request not to put on the brakes.
The timing also counts. Days earlier, Dario Amodei, CEO of Anthropic, the company that makes the Claude assistant, published a text defending slowing the pace of AI advancement, with independent evaluators inside companies and agreements between countries.
Trump responded on Truth Social, on September 19, with a post in which he calls attacks against AI a hoax and says they are the work of Democrats.
The UN renaming came three days later. Anyone wanting the complete picture of who defends rules and who is against has the article about AI regulation and the fight between Trump, Nvidia, OpenAI and Anthropic.
There is also an effect outside the United States. Experts interviewed by Ars Technica assess that the obsession with "winning the AI race" against China hinders cooperation to make AI safe, avoiding accidents and misuse, between the two countries that most develop the technology. "Whoever wins in super intelligence wins" is exactly this idea, with the new name.
Jensen Huang and Trump: Nvidia's ally
Jensen Huang, CEO of Nvidia, became the industry's strongest voice in favor of Trump's line.
Nvidia manufactures the chips (the parts that do the calculations, like a car's engine) used to train and run almost all advanced AI in the world. Running is answering users day to day. Training is the phase when AI "studies" billions of texts, like a student before an exam, and consumes thousands of these chips. Therefore, any brake on the industry is a brake on its sales.
The scene that marked this alliance was at the same All-In Summit. With Huang on stage, according to TechCrunch's report and the event recording, Trump called his cell phone, and Huang put the call on speakerphone for the audience. Trump said that those asking for deceleration play into political interests and perhaps China's, and added: "We're not going to let that happen. It's a hoax". Huang replied: "You're right. We're not going to let that happen, sir".
In the same conversation, Trump said it is necessary to act "with prudence", but that does not mean stopping the industry. "I'm with you all the way", he told Huang, according to the event recording.
On the Equity podcast, TechCrunch journalists summarized the relationship: Huang has become one of the most useful bridges between the industry and government, and his and Trump's interests are aligned. One of the hosts said the scene seemed rehearsed and made clear how much Nvidia gains from an AI that advances without brakes.
In one sentence: the new name and the alliance with Huang tell the same story: the American government wants AI big, fast and without rules.
What this changes for you
In practice, almost nothing changes today. ChatGPT, Claude and Gemini remain the same, and no one needs to change the name of anything. What changes is the vocabulary that will appear in news, speeches and advertisements. Knowing what each word means helps you not get carried away by it.
| Your profile | What to do |
|---|---|
| Uses AI at work or in studies | Nothing changes in the tools. If a product sells itself as "superintelligent", be suspicious: by definition, that does not yet exist. |
| Student or teacher | Use the confusion as a lesson: "artificial" (made by humans), AGI and superintelligence are different concepts, with dates and authors. |
| Follows politics and technology | Look at what comes with the name: refusal of international rules and surveillance only by the Department of Justice. |
| Works with companies or government | American documents may start using "super intelligence" (SI). In contracts and translations, treat it as synonymous with AI, not as new technology. |
A warning for those who translate or write: the technical concept from Bostrom is written "superintelligence", one word. "Super intelligence", separated, is the literal translation of the name Trump proposed, and that is how it has appeared in the news. Worth separating the two: one is a scientific concept, the other is a political label.
To follow the subject beyond the renaming, the artificial intelligence page at Kalkazu gathers guides and news from the industry.
Frequently asked questions
Did Trump really change the name of artificial intelligence?
He announced the change in the UN speech on September 22, 2026, and the White House published the sentence. But no decree has been issued so far. Even with a decree, it would only apply to U.S. federal government documents, not to the rest of the world.
What does SI mean?
SI is the acronym for super intelligence, the name Trump proposed instead of AI (artificial intelligence). In the official summary of the speech, the White House wrote the name in full, "Super Intelligence".
Is ChatGPT already a superintelligence?
No, by the technical definition. Superintelligence would be an AI much better than the best humans at practically all areas at the same time. ChatGPT is very good at several tasks and still makes factual errors, invents data and depends on people to be useful.
Will the world have to use the term super intelligence?
No. Even if a decree is issued, it only applies to the United States government, as happened with the "Gulf of America". Everywhere else, laws, companies and the press remain free to call the technology artificial intelligence, which is the term used worldwide.
Why does Jensen Huang support Trump on AI regulation?
Nvidia sells the chips that power AI, so the faster the industry grows, the more it sells. Huang and Trump agree that slowing down would be a mistake and that risk warnings are exaggerated. Critics note that this alignment directly benefits his company.
A name bigger than the technology
Swapping "artificial" for "super" does not change any chip, any model or any answer from your AI assistant. It changes the packaging, and the packaging tells a story: that the technology has already arrived and does not need brakes. Next time you hear "super intelligence", remember that the term describes a machine no one has built yet.
