The September 29 Executive Order remaps executive-branch vocabulary, not the statute, and the collision with the technical meaning of superintelligence is the governance story that matters.
HAIA Agents · Responsible AI · September 30, 2026
Federal agencies now face a vocabulary order that treats "Artificial Intelligence" as a term they should stop acknowledging in non-statutory work. On September 29, 2026, President Donald J. Trump signed the Executive Order “Inaugurating the Era of Super Intelligence”. The order directs executive departments and agencies, to the maximum extent permitted by law, to use "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI." The directive covers official correspondence, public communications, websites, reports, policy documents, and other non-statutory materials. The order also states that the executive branch will not acknowledge usage of the older terms in applicable settings.
That is branding with binding force inside the executive branch, and it is not a silent rewrite of federal statute. Section 2(b) of the order says nothing requires alteration of previously issued regulations, Presidential actions, contracts, grants, or historical documents. For implementation, Section 3(a) defines "Super Intelligence" and "SI" as the same technologies and systems already covered by "artificial intelligence" under section 9401(3) of title 15, United States Code. That mapping holds until later Presidential action or an Act of Congress supersedes it. The Assistant to the President for Science and Technology has sixty days to propose legislative language for a federal SI definition. That proposal must assess whether the definition should modify or replace the statutory AI definition, and it must recommend conforming amendments.
The practical question for Responsible AI operators is narrower than the political fight over the label. When government language, industry standards, and research vocabulary diverge, compliance teams, communications staff, and model-risk owners inherit dual lexicons that can hide disagreement rather than clarify it.
What the September timeline actually changed
The rename did not begin with the Executive Order. On September 22, 2026, Trump told the United Nations General Assembly that the word "artificial" makes intelligence sound fake, and that United States documents should use "super intelligence" or "S.I." going forward. UN News recorded the competitiveness line that followed: "Whoever wins superintelligence wins." Within hours, reporting from The Hill and related outlets described a State Department Bureau of International Organization Affairs email on the same subject. That message told diplomats to replace "Artificial Intelligence" with "Super Intelligence" and to use "SI" in remarks, positions, and press elements.
Crowdsourcing preceded the UN speech, and Newsweek reported a Truth Social poll with more than 50,000 responses in which roughly 65 percent preferred "super intelligence" over "superior intelligence." BBC coverage notes a wider set of poll options that included extreme and supreme variants, with super intelligence as the winner.
On September 29, after a White House lunch with leading technology executives, the formal Executive Order locked the vocabulary rule into executive-branch practice. CNBC reporting draws a hard line operators should keep about the lunch documents. The "White House Accord on Super Intelligence," described as a joint commitment on frontier responsibilities and labeled "morally binding" by Trump, is a separate voluntary document. That accord, as described in the coverage, does not declare an industry name change even though its title uses the new federal phrase. Signatories shown in the Truth Social post Trump shared include Sundar Pichai, Elon Musk, Dario Amodei, Mark Zuckerberg, Greg Brockman, and Jensen Huang, according to CNBC.
House Speaker Mike Johnson, speaking to reporters that day, framed the accord as voluntary commitments with internal controls and layers of review. Fox Business quotes him saying safety and innovation aren’t mutually exclusive pursuits. Meta CEO Mark Zuckerberg wrote on X that major American lab leaders committing to internal controls he called "robust," plus multiple layers of audits, was a "significant positive step."
Why the White House wants Super Intelligence as the federal label
The White House Fact Sheet presents the change as recognition of capability, not cosmetics. It argues that Super Intelligence conveys the true capabilities of systems being developed today, unlocking creativity and empowering Americans, and that "Artificial Intelligence" implies imitation or automation of human intelligence. The Executive Order’s purpose section makes the same claim in denser form. It says frontier systems "do much more than imitate or automate discrete aspects of human intelligence." As those capabilities improve, the text continues, the systems "increasingly represent not merely artificial intelligence, but a new era of Super Intelligence."
The policy package around the label is already thick. The Fact Sheet points to more than $5 billion announced for the Genesis Mission, America’s AI Action Plan from July 2025 with more than ninety federal actions, and a March 2026 National Super Intelligence Policy Framework with legislative recommendations. In that framing, the rename keeps agencies focused on expanding capability and application rather than on a term the administration treats as outdated.
Trump’s own public case mixes brand preference with geopolitical competition in ways operators can’t ignore. At the UN he rejected what he called a globalist scheme to control the technology, said the United States would encourage Super Intelligence, and said law enforcement bodies would rein things in if required. That pairing of acceleration language with a residual enforcement claim is the live policy tension operators inherit when they have to interpret federal SI documents against older AI statutes that still use the statutory term.
Who is lining up behind the SI vocabulary
Administration figures moved first. BBC reporting quotes U.S. Chief Technology Officer Ethan Klein adopting "Super Intelligence (SI)" language immediately after the UN speech. UN ambassador Michael Waltz said "it’s now super intelligence" and called the president "absolutely right." The State Department email after the General Assembly shows an agency already translating speech into drafting instructions before the September 29 order existed.
Industry alignment is partial and instrumental rather than definitional, and that distinction matters for anyone tracking standards. The tech executives who attended the White House lunch participated in a voluntary accord that Johnson and Zuckerberg describe in control-and-audit terms. University of Chicago marketing professor Jean-Pierre Dube told Newsweek the rebrand is unlikely to take off unless tech leaders adopt the term. He then added a practical caveat: if CEOs anticipate fast-tracked infrastructure or land permitting for data centers, they might adopt SI language for instrumental reasons. Using SI in a White House meeting and accepting SI as the research community’s term of art are different acts.
Support for the rename also sits inside a broader administration preference for industry self-policing over new mandatory guardrails. CNBC reports Trump dismissing dire warnings as a "hoax" even as some executives at the table have publicly raised safety concerns in other forums. Johnson’s "statement of principles" language and Trump’s "morally binding" description of the accord are the affirmative case for voluntary discipline under the new label.
Why critics say the rename creates risk rather than clarity
The strongest technical objection is lexical collision. In research and safety discourse, superintelligence already names a hypothetical intellect that greatly exceeds human cognitive performance across domains. That usage is long associated with Nick Bostrom’s work and is still treated as speculative. Digital ethics lecturer Simon Coghlan of the University of Melbourne told the BBC the rebrand is "misleading" and that he doubts SI will stick because it exaggerates current capacities.
Computer scientist Kokil Jaidka said the new name "borrows the prestige of a term researchers reserve for a much bigger leap." National University of Singapore lecturer Ben Leong predicted professionals will want to reserve SI for a higher capability tier than systems available today.
Marketing scholars amplify the public-trust worry, and Dube told Newsweek that "super intelligence" may spark greater distrust than "artificial intelligence" because it points toward technology people fear they can’t control. Boston University’s Susan Fournier argued the label implies intelligence superior in scale and power, and therefore more threatening. For people who know the technical definition, she said, it heightens science-fiction fears while making safety regulation harder to imagine. Fournier’s forecast is blunt: she would bet on the tech industry winning the label war.
Political criticism treats the rename as a substitute for substantive risk policy. Sen. Mark R. Warner, vice chairman of the Senate Intelligence Committee, said in a statement reported by CNBC that companies building the most powerful systems are warning that technology is advancing faster than safeguards. Warner said the president’s response was "to rename it and tell the companies developing it to regulate themselves." Changing what the government calls artificial intelligence, Warner said, "does nothing to address the very real risks that come with increasingly powerful AI systems."
Public sentiment gives that criticism a measurement backdrop. A Quinnipiac University poll conducted September 24 to 27, released the day of the Executive Order and reported by CNBC, found 25 percent approving and 58 percent disapproving of Trump’s handling of AI. That is a weaker approval figure than several other issues in the same survey. The rename arrives inside that approval gap rather than after it closes.
International coolness shows up in smaller reactions, and Canada’s minister of AI, Evan Solomon, told the BBC that Canadians are "not enthusiastic, at all, about renaming things," in an apparent reference to other Trump geographic rebrands.
What Responsible AI operators should change in the next sixty days
The Factics move is concrete. Treat the Executive Order as a verified fact about executive-branch communication rules, and don’t treat it as proof that the technical category called AI has been retired. The tactic is dual-lexicon hygiene built around a controlled vocabulary table. That table should map statutory AI under 15 USC 9401(3), federal SI under the September 29 order, and research "superintelligence" as three labeled concepts with different evidence strength and different audiences. The KPI is error rate in external filings and client briefings over the next two reporting cycles. Count how often a document uses SI where a statute, standard, contract, or research claim still requires AI or the Bostrom-class term, and drive that mismatch toward zero before the APST’s sixty-day proposal lands.
Communications and SEO teams face a parallel workflow problem as publication channels diverge. Federal press material, grant language, and agency websites will begin saying SI while Google results, standards bodies, academic libraries, and most commercial documentation still index AI. Operators who publish under Responsible AI disclosure rules already treat attribution as a governed claim. The same discipline belongs on terminology, so quote the White House when the claim is about federal vocabulary, quote the statute when the claim is about legal scope, and quote the research literature when the claim is about systems that exceed human-level performance across domains.
Checkpoint-Based Governance supplies the authority test behind that hygiene, because a rename can change what an agency writes on a webpage without relocating binding human judgment. If a lab, vendor, or agency treats SI branding as evidence that frontier risk has been solved, the checkpoint has been skipped, and new language on a federal page does not transfer accountability for outcomes.
Watch the next layer rather than the slogan after the sixty-day definition clock starts running. The unresolved items include the APST’s legislative proposal and whether Congress ever moves statutory references. Operators should also watch whether industry standards bodies adopt SI as anything more than courtesy language in Washington, and whether public trust improves or erodes when federal documents call today’s systems Super Intelligence while safety researchers reserve that phrase for a different class of system. Those four watchpoints decide whether the rename clarifies policy language or multiplies it.
Frequently Asked Questions
What did Trump’s Super Intelligence Executive Order change?
On September 29, 2026, the order directed executive agencies, to the maximum extent permitted by law, to use Super Intelligence and SI instead of Artificial Intelligence and AI in non-statutory communications and documents. It does not require rewriting prior regulations, contracts, grants, or historical documents.
Is Super Intelligence a new legal definition of AI?
For now, no. Section 3 maps SI to the existing statutory artificial intelligence definition in 15 U.S.C. section 9401(3). The Assistant to the President for Science and Technology has sixty days to propose legislative language for a federal SI definition.
How is federal SI different from research superintelligence?
Federal SI under the order currently covers the same systems already called AI in statute. In research, superintelligence usually names a hypothetical intellect that greatly exceeds human performance across domains, a usage associated with Nick Bostrom and still treated as speculative.
Did tech CEOs rename AI when they signed the White House Accord?
Coverage from CNBC and Fox Business treats the White House Accord on Super Intelligence as a separate voluntary commitment on frontier responsibilities. Trump called it morally binding, while Speaker Johnson described voluntary principles and internal controls. The accord title uses SI language without declaring an industry-wide name change.
What should Responsible AI teams do next?
Build a dual-lexicon vocabulary table that separates statutory AI, federal SI, and research superintelligence, then measure terminology mismatch errors in filings and briefings before the sixty-day APST proposal arrives. Treat branding as communication policy rather than proof that frontier risk controls are complete.
Where can readers verify the primary documents?
The White House published the full Executive Order and a Fact Sheet on September 29, 2026. UN News and the Miller Center transcript cover the September 22 UN General Assembly announcement that preceded the order.
#AIgenerated
Leave a Reply
You must be logged in to post a comment.