AI Is Now 'SI': What Trump's Renaming Actually Means
On September 22, 2026, the US government began replacing 'artificial intelligence' with 'super intelligence' (SI). A September 29 executive order made it official policy, and an October 4 'Super Intelligence Force' gave it an institution. Full timeline, the ASI terminology collision, and Federal Register adoption data.
On this page
The short answer
Updated October 16, 2026 — now covering the September 29 executive order, the Federal Register publication, and the Super Intelligence Force.
On September 22, 2026, President Trump told the UN General Assembly that all US government documents would stop saying “artificial intelligence” and start saying “super intelligence” — SI. Within hours, the State Department’s Bureau of International Organization Affairs ordered its staff to comply. The renaming is real, but its adoption is not yet universal, and it collides with an existing technical term that means something much more specific.
For companies selling to the US government, for analysts, and for anyone running search-driven content in this space: the vocabulary of the largest AI buyer on Earth is now officially in motion.
The timeline
| Date | Event |
|---|---|
| Sep 21, 2026 | Trump posts on Truth Social: “I will only encourage AI or, SI (SUPER INTELLIGENCE)!” — first public use of the rebrand (The Hill) |
| Sep 22, 2026 | In his UN General Assembly address, Trump announces the change: “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’” (Breaking Defense, Axios) |
| Sep 22, 2026 | The stated reason: “The use of the word artificial… makes it sound fake and it is not fake. It’s actually amazing” (Axios) |
| Sep 22, 2026 | State Department email, subject line “Change in Nomenclature AI to SI,” instructs the Bureau of International Organization Affairs: “We will be using ‘SI’ in all remarks, positions and press elements moving forward” (AP) |
| Sep 23, 2026 | US representatives use “SI” at a UN Security Council meeting on AI (AP) |
| Sep 29, 2026 | Trump signs the Executive Order “Inaugurating the Era of Super Intelligence,” directing that federal usage of “artificial intelligence” and “AI” be replaced by “Super Intelligence” and “SI” (White House, Fact Sheet) |
| Oct 2, 2026 | The EO is published in the Federal Register. Legally, “Super Intelligence” and “SI” now mean exactly what “artificial intelligence” meant in 15 U.S.C. § 9401(3) — same definition, new words (Federal Register) |
| Oct 3, 2026 | Director of National Intelligence Jay Clayton is named the administration’s AI czar, leading a new task force with 120 days to report on the risks and opportunities of the technology and the federal government’s responsibilities (WSJ, CNBC) |
| Oct 4, 2026 | The “Super Intelligence Force” — the task force itself — is announced (The Hill) |
| Oct 6, 2026 | Politico publishes the SIF charter: the task force is to “coordinate the effort of the Federal Government to ensure that America continues to lead the World in Super Intelligence” (Politico) |
| Oct 13, 2026 | Task force member Scott Kupor (a16z) says companies will not get a “free pass” on safety, while cautioning against rules that inhibit US innovation (Politico, PYMNTS) |
Why this is policy, not just words
The renaming was not a stray remark. It arrived inside a broader repositioning: in the same speech, Trump said the US “totally rejects any attempt to construct a globalist scheme to control” AI, explicitly framing AI safety advocates in the same terms he uses for climate warnings, and promising “we will only encourage super intelligence… We’re going to encourage it, not rein it in” (Axios, AFP).
The vocabulary does consistent work here. “Artificial” carries connotations of fake, synthetic, lesser-than. “Super” carries connotations of dominance — which matches a must-win race framing against China. Terminology that sounds like a product demo is easier to encourage than terminology that sounds like a risk category. Meanwhile, leaders of more than 20 mostly European countries were pushing in the opposite direction at the same assembly, calling for binding safety measures (AFP).
There is also a domestic administrative pattern at play: this administration has previously renamed the Department of Defense as the “Department of War” in messaging, even where statute still says otherwise (Breaking Defense). Unlike that case, replacing “artificial” with “super” creates no legal complications — which makes this particular rename cheap to execute.
The terminology collision
Here is the problem the rebrand creates. “Superintelligence” is not a free word. For years it has had a precise meaning in AI research and in the industry’s own communications:
| Term | Established meaning | Status |
|---|---|---|
| AI — artificial intelligence | The broad field and its current products | Being retired in US official use |
| AGI — artificial general intelligence | Hypothetical human-level AI across domains | Unchanged |
| ASI — artificial super intelligence, superintelligence | Hypothetical AI smarter than the best humans in every domain; does not exist yet | Now politically borrowed |
| SI — super intelligence (new US usage) | What the US government previously called AI | Born Sep 22, 2026 |
As Breaking Defense noted after the speech, it is “not entirely clear whether Trump is now endorsing superhuman machine minds, or simply saying AI as it exists today is already ‘super’” (Breaking Defense). That ambiguity will propagate. When a federal document says “SI,” a reader can no longer tell whether it means a chatbot, an agent pipeline, or a recursively self-improving system. Regulatory language that loses precision is language that becomes harder to enforce.
For research organizations and frontier labs, this is a forced disambiguation problem: their safety documents use “superintelligence” in the strict sense, while US government communications will use “SI” in the loose sense. Watch for the emergence of qualifiers — “true superintelligence,” “ASI (technical sense)” — as the two usages grind against each other.
What it means in practice
For federal contractors and vendors. Where government documents go, procurement vocabulary follows. The Department of War precedent suggests adoption in messaging well ahead of statute. The open question AP flagged on September 23 — whether the “AI Force” would become the “SI Force” — was answered on October 4: the Super Intelligence Force is real, chaired by AI czar Jay Clayton (The Hill). Contractors rewriting capability statements should track agency-by-agency adoption rather than assume a single switch date. For a breakdown of the task force itself, see our companion piece: The Super Intelligence Force, explained.
For analysts and strategists. Treat “SI” as a new politically-loaded synonym, not a new technology category. If your dashboards and keyword taxonomies key off “AI,” add “SI” and “super intelligence” as monitored variants now; if the term gains traction in federal and then media usage, early movers get the clean search positions. If it stalls at the State Department, you have lost nothing.
For the rest of the world. The EU AI Act, the Bletchley process successors, and the 20+ European leaders calling for binding safety measures all remain anchored to “AI” (AFP). A transatlantic vocabulary split is now plausible: US official documents saying SI while European regulators say AI — referring to the same systems. In diplomatic contexts, that is not cosmetic. Shared terminology is a precondition for shared rules.
How to tell whether SI sticks
Three observable signals, in rough order of lead time:
- Other agencies. The State Department’s international organizations bureau was a single bureau, not the government (AP). The September 29 executive order changed that — every agency is now directed to make the switch (White House). We can measure adoption directly in the Federal Register, whose API supports full-text search. Documents mentioning each term, by month:
| Month | ”artificial intelligence" | "Super Intelligence” |
|---|---|---|
| May 2026 | 39 | 0 |
| Jun 2026 | 23 | 0 |
| Jul 2026 | 39 | 0 |
| Aug 2026 | 34 | 0 |
| Sep 2026 | 30 | 0 |
| Oct 2026 (through Oct 16) | 13 | 3 |
Three documents in the first half of October, from zero in every prior month of 2026. Slow, from zero — which is how administrative vocabulary actually starts. We will re-run this query monthly and update the table.
- Contract language. The first SI-labeled solicitation or program name in federal procurement is the moment the term acquires budget gravity. None yet — the Federal Register mentions so far are policy documents, not procurements.
- Style guides. When wire services and major outlets decide whether to write “SI (super intelligence, formerly AI)” or to keep “AI” with a note — that is the mass-adoption gate. The AP and Axios stories cited above still say “AI” in their own voice, with “SI” in quotes — the transition is not done on this front either.
We will track these on Revolutio as part of our disruption monitoring. Fittingly, our own domain ends in .si — a coincidence we had no way to plan for in 2026’s naming lottery, and one we intend to enjoy.
Bottom line
A government does not rename a technology it intends to restrain. The AI-to-SI shift is a signal about the direction of US AI policy — encouragement over restraint — wrapped in a terminology collision that will confuse precisely the discussions about limits that the rename’s author rejects. The word “artificial” made AI sound fake; the word “super” makes oversight sound optional. Both framings are doing political work, and neither describes the systems themselves.
The disruption story we track — jobs, industries, measured gains — does not change because of a rebrand. But the language used by the world’s largest AI funder just split from the language used by everyone researching its risks. That split is worth monitoring on its own terms.