The glossary
The building blocks of the Super Intelligence story — the order, the Accord, the people, the fine print — each defined briefly, with dates where they matter.
Term in use since the 1950s
The term the September 2026 executive order retires from federal executive-branch usage. Coined in the 1950s around the Dartmouth research project, it was the standard name for the field for seventy years — and it still is, everywhere outside the U.S. executive branch: in law, in industry, in academia, and in the EU’s Artificial Intelligence Act.
September 29, 2026
A one-page voluntary agreement signed at the White House by President Trump and the leaders of Google, Anthropic, Meta, OpenAI, xAI, and NVIDIA. It commits the companies to four layers of safeguards: robust internal controls on model capabilities and alignment; a dedicated internal team to verify those controls work; an independent external auditor or evaluator; and an independent committee of each company’s board of directors to oversee the process. The controls cover risks including cybersecurity, biosecurity, and chemical threats, and are meant to ensure models do not “hack or access technical systems in unintended ways.” The Accord is voluntary — the president described it as “morally binding” — and leaves open that the measures could later be written into law.
Launched September 29, 2026
A federal website launched the morning of the signing, presented as a “one-stop shop” for Americans navigating the federal government. The president used its launch event to announce the renaming.
60-day deadline: ≈ November 28, 2026
The White House science adviser. The executive order gives the APST 60 days to submit proposed legislative language establishing a formal federal definition of “Super Intelligence,” including whether it should modify or expand the existing statutory definition of AI, what conforming amendments to current laws would be needed, and what further executive action the definition would require.
Signed September 29, 2026
The order that renames AI to SI across the federal executive branch. Its key provisions: agencies must use “Super Intelligence” and “SI” in place of “artificial intelligence” and “AI” in official correspondence, public communications, websites, reports, policy documents, and other non-statutory documents; it does not require rewriting previously issued regulations, presidential actions, contracts, grants, or historical documents; and for legal purposes “SI” means exactly what the statute already defines as AI (15 U.S.C. § 9401(3)) — until Congress or a later presidential action says otherwise. The order’s stated rationale: modern systems “do much more than imitate or automate discrete aspects of human intelligence,” so federal terminology should reflect their “transformative capabilities.”
Term in policy use; June 2026 EO
The most advanced AI systems — the ones at the cutting edge of capability. The Accord’s commitments are aimed specifically at companies training and deploying frontier models. In June 2026, a separate executive order had already set up voluntary pre-release vetting of frontier models and a federal clearinghouse for AI-related cybersecurity risks, stopping short of mandatory licensing.
September 29, 2026
The president’s own description of the White House Accord: “It’s almost like a constitution in a way… I think it’s morally binding.” The phrase captures the Accord’s status — a serious public commitment by the signatories, but not a law, regulation, or contract, with no legal penalties for non-compliance.
As used in the September 29, 2026 order
The category of government writing the rename applies to: correspondence, public communications, websites, reports, policy documents, and similar materials. Statutes passed by Congress, existing regulations, contracts, and grants are explicitly excluded — only Congress can change the words in the law itself.
Ongoing
The White House office that advises the president on science and technology. Its director serves as the Assistant to the President for Science and Technology — the official tasked with developing the proposed federal definition of “Super Intelligence” within 60 days of the order.
Federal usage from September 29, 2026
The replacement term for “artificial intelligence” / “AI” in U.S. executive-branch usage. The president’s argument: “artificial” makes the technology sound fake, while these systems are “very powerful” and “very brilliant.” Legally, it changes nothing — SI covers the same technologies the statute defines as AI. Note the spacing: the government writes “Super Intelligence” (two words); researchers have long used the single word superintelligence to mean something different (see below).
Research term; popularized 2014
In AI research and philosophy, “superintelligence” means a hypothetical system that surpasses human intelligence across the board — popularized by Nick Bostrom’s 2014 book Superintelligence. This is not what the executive order means by “Super Intelligence.” The order’s term is a rebrand of today’s AI; the research term describes a possible future system. The overlap in wording has already caused confusion, which is one reason this glossary exists.
September 2026
The runner-up name. Before the announcement, the president ran a poll on Truth Social asking followers to choose between “super intelligence” and “superior intelligence.” About 65% voted for “super intelligence,” which became the official pick.
September 29, 2026
The administration’s stated approach: rather than new mandatory rules, the leading companies police themselves under the Accord’s four layers of controls, with the possibility of laws later. “I think I’m seeing tremendous self-policing,” the president told reporters after the signing. “They understand that they have to self-police.”
September 2026, before the order
Earlier in September 2026, the State Department’s international organisations bureau instructed diplomats to use “super intelligence” or “SI” in place of “artificial intelligence” — a preview of the government-wide rename that followed on September 29.
Surge from September 22, 2026
.si is the country-code top-level domain of Slovenia — a Central European country between Italy, Austria, Hungary, and Croatia. Anyone in the world can register one. After the president first used “super intelligence (SI)” at the UN on September 22, .si registrations overtook .ai within about 25 hours (Netcraft), running near 3,700 new domains a day versus about 3,500 in all of August, pushing the total past 220,000. Slovenia’s registry noted the surge and publicly advised trademark holders to secure their .si names. More than half of the speculative registrations were quickly listed for resale on the domain aftermarket.
September 2026
The public vote that picked the name: followers chose “super intelligence” over “superior intelligence” by about 65% to 35%.
September 22, 2026
Where the rebrand debuted publicly. Addressing the UN General Assembly, the president said the United States would stop calling the technology “artificial intelligence” in official documents because the word “artificial” makes it sound fake: “Welcome to the new world of super intelligence — SI.” The .si domain surge began within hours.
September 29, 2026
The seven signatures on the White House Accord on Super Intelligence: President Donald J. Trump; Sundar Pichai (Google); Dario Amodei (Anthropic); Mark Zuckerberg (Meta); Greg Brockman (OpenAI); Elon Musk (xAI); and Jensen Huang (NVIDIA).
Enacted definition; referenced September 29, 2026
The section of federal law that defines “artificial intelligence.” The executive order explicitly ties “Super Intelligence” to this existing definition — so while the label changes in executive-branch documents, the legal scope of what counts as AI/SI does not, unless Congress acts.
Definitions current as of September 30, 2026. Spot something out of date? This page is updated as the story develops.