Short answer: the technology did not change, only the label in US government documents. On 29 September 2026 President Trump signed an executive order telling federal agencies to write “Super Intelligence” (SI) instead of “Artificial Intelligence” (AI). It binds only the US executive branch, defines SI exactly as the existing legal definition of AI, and leaves laws, contracts and private companies’ product names untouched. For most readers it changes news vocabulary, not what these systems can do.
Key facts: dates and sources
- 22 September 2026: at the UN General Assembly, Trump said he would rename artificial intelligence because “artificial” makes it sound fake (AP via ABC News).
- The same week: a State Department bureau told staff to use SI in remarks and press material (AP).
- 29 September 2026: Executive Order 14434, “Inaugurating the Era of Super Intelligence” (order text, White House fact sheet).
- 30 September 2026: California’s governor reportedly told state agencies to keep saying AI (IBTimes).
- 1 October 2026: YouGov found 53% prefer “artificial intelligence” and 9% “super intelligence” (Axios).
- 4 October 2026: Elon Musk said SpaceXAI will be renamed SpaceXSI, with no date (Not a Tesla App).
- 28 November 2026: deadline for proposed legislative language defining SI (Freshfields).
What was said at the UN
On 22 September, speaking at the UN General Assembly in New York, Trump said he intended to change the name of artificial intelligence. His argument was about language, not engineering: in his view “artificial” signals something fake, while the technology is real and powerful. He added that US government documents would use “super intelligence” from then on. Formal steps followed within days.
The State Department moved first
AP reported that Michael Drager, a deputy assistant secretary in the Bureau of International Organization Affairs, emailed staff under the subject line “Change in Nomenclature AI to SI”. The bureau would use SI in all remarks, positions and press material. Because that bureau oversees US missions to the UN and other multilateral bodies, the instruction reaches far. It was unclear whether other bureaus would follow.
What Executive Order 14434 says
The order, signed on 29 September and published in the Federal Register on 2 October, tells departments and agencies to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official correspondence, public communications, websites, reports and policy documents, “to the maximum extent permitted by law”. It sets no penalty for anyone who keeps writing AI, and section 4 states that it creates no right enforceable in court.
The definition: SI is whatever AI already was
Section 3(a) says Super Intelligence and SI mean the technologies and systems covered by “artificial intelligence” as already defined in US law, at 15 U.S.C. 9401(3). In plain terms there is no new capability threshold, no test and no new category of technology. Anything that counted as AI the day before the order counts as SI the day after, nothing more and nothing less.
What the order does not change
Section 2(b) says agencies do not have to alter previously issued regulations, presidential actions, contracts, grants or other historical documents. Freshfields, which analysed the order, stresses that it targets non-statutory documents and that existing statutes still say AI. It governs the wording of government paperwork, not the systems themselves.
A new definition is on the way
Under section 3(b), within 60 days the Assistant to the President for Science and Technology, who heads the White House Office of Science and Technology Policy (OSTP), must send the president proposed legislative language for a federal definition of SI. Freshfields puts the date at 28 November 2026. That is a proposal, not a law. Replacing the statutory definition would take an Act of Congress, and until then the old definition stands.
Who is bound and who is not
The order speaks to the federal executive branch: departments, agencies and official websites. It does not bind Congress, the courts, US states, universities, newsrooms or private companies, let alone foreign governments. A company in Europe, Israel or anywhere else has nothing to change. Even inside the US, a business that keeps writing AI breaks no law.
What the big AI companies did
At the signing, executives from OpenAI, Anthropic, Google, Nvidia, Meta and SpaceXAI stood beside the president and signed a pledge with “super intelligence” in its title. The SF Standard reported there was no explicit agreement to rebrand their products, and Freshfields noted the pledge creates no legally enforceable obligations. OpenAI, Google and Anthropic still describe their products as AI. Meta’s unit called Meta Superintelligence Labs dates from 2025, long before the order.
Musk: SpaceXAI to become SpaceXSI
On 4 October Elon Musk replied to a user on X that SpaceXAI would be renamed SpaceXSI and that SI is the better term. He gave no date, and reports from those days said the official accounts and domains still read SpaceXAI. It would not be the unit’s first rename: after SpaceX acquired xAI, it became a division called SpaceXAI.
California sticks with AI
A day after the signing, on 30 September, IBTimes reported that California Governor Gavin Newsom ordered state agencies to keep using “Artificial Intelligence” and “AI”. That matters in practice: many of the largest AI companies are based in California, and the state writes its own oversight laws. So the US itself will run two vocabularies side by side for a while.
What the American public thinks
A YouGov poll reported by Axios in the week of the signing found that 53% of US respondents considered “artificial intelligence” the better term, against 9% who preferred “super intelligence”. But it shows the new label has not caught on outside government, and most media and companies keep writing AI. Expect to see both terms in the news for a long time.
Even the White House has not fully switched
Axios reported on 8 October that Trump wrote on Truth Social that the White House considers anyone using “Artificial Intelligence” to be “THE ENEMY”. The same report noted the White House itself had not finished the switch: the AI.gov homepage shows “SI.gov” in large type, but the address is still AI.gov, and the OSTP website still says “artificial intelligence”.
What superintelligence means in research
Before the order, “superintelligence” already had an established meaning. Philosopher Nick Bostrom, in his 2014 book Superintelligence, used it for an intellect that greatly exceeds human cognitive performance in virtually every domain. Much of the debate about safety, oversight and the pace of development revolves around that idea.
New label, old meaning
This is where confusion starts. In research, superintelligence describes a future, theoretical capability. In the order, SI is a new name for everything already called AI: a spam filter, a chatbot, an image recogniser. Same word, different things. So when a US government document says “SI”, it is not claiming superintelligence in the research sense, and Axios noted that Trump does not appear to mean that kind of system. Always ask who wrote it and in what context.
Why the word feels unsettling, and why panic is not warranted
“Superintelligence” sounds like science fiction, so it is natural to feel uneasy when it suddenly fills the headlines. But the order does not follow a technical breakthrough and arrives with no announcement of new capability. It is branding and terminology. The systems you use today behave exactly as they did the week before the speech. Concern or excitement should rest on what systems do, not on what they are called.
Are today’s systems superintelligent?
By the research definition, no, and researchers do not agree that they are. Language models and computer vision systems excel at specific tasks, sometimes beyond people, but they also make mistakes, invent facts and depend on their training data. Whether and when a system will exceed humans in almost every domain is contested, and expert estimates vary widely. Treat anyone who presents the new label as proof that it has already happened with caution.
What changes for readers outside the US
Legally, almost nothing. The order is a document of the US executive branch and has no force in the EU, the UK, Israel or elsewhere. Local laws, regulators, the terms of the services you use and the contracts you signed stay as they are. Bidders on new federal tenders may meet SI in the paperwork, and that is roughly the limit of direct impact.
Reading the headlines from now on
The most visible change will be in the news. Statements from the White House, the State Department and US diplomats will say SI, and the media will quote them. A simple rule of thumb: when the term comes from the US government, it almost always means the same AI you already know. When it comes from a research paper or from the heads of AI labs, check whether it refers to a future capability.
Scams will ride the new buzzword
Every hot term becomes bait. It happened with crypto and with “AI”: investment offers with guaranteed returns, a “trading bot” that doubles your money, an expensive course promising “the secret of SI”. A new name does not make a product smarter, and it may signal that someone is exploiting the confusion. The rules that protect you from other scams apply here too.
How to spot a suspicious “SI” offer
Warning signs worth stopping for: promised fixed or “safe” profits; time pressure such as “today only”; requests to send money to a crypto wallet or an overseas account; and the names of Trump, Musk or the White House used to borrow credibility. Before investing, check that the firm is licensed by your national financial regulator. If children or older relatives receive such messages, talk about it; our guide to spotting online scams (in Hebrew) works for adults too.
For businesses: do you need to change anything?
In most cases, no. Customers search for “AI”, so keep that word on your website, in proposals and in marketing. If you deal with US federal agencies, use SI in new documents for them without touching existing contracts. To help search engines and AI assistants understand your business, describe clearly what your system does rather than what it is called; see what GEO is (in Hebrew). If you are weighing an AI agent for your business, the label changes nothing about what it can do, and a personal AI consultation can test that against your case.
What to watch in the coming months
Three things. First, the federal definition due to the president by 28 November, and whether Congress turns it into law. Second, whether more companies, beyond Musk’s statement, rename products or units. Third, whether other governments adopt the term in their official statements. Until then this is mainly a change of wording in US government documents.
Bottom line
As of October 2026, SI and AI are two names for the same technology. Trump’s order binds only US federal agencies, changes no laws or contracts, and defines SI by the existing definition of AI. Superintelligence in the research sense, an ability beyond humans in almost every field, was not achieved by decree. Read headlines critically, and be wary of anyone selling something with the new word. For more background, see our guide to AI for business (in Hebrew).