On September 29, the White House issued Executive Order 14434, “Inaugurating the Era of Super Intelligence.” It directs every executive department and agency, “to the maximum extent permitted by law,” to use Super Intelligence and SI in place of Artificial Intelligence and AI in correspondence, public communications, websites, reports and policy documents. It also says the executive branch “will not acknowledge the usage of ‘Artificial Intelligence’ and ‘AI’ in any applicable setting.”
Now read Section 3.
For the purposes of the order, Super Intelligence means “the technologies and systems encompassed by the term ‘artificial intelligence’” as defined in 15 U.S.C. § 9401(3). That is the National Artificial Intelligence Initiative Act of 2020: “a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments.” The fraud model in your payments stack meets that definition. So does the spam filter. As of this month, both are Super Intelligence in federal usage.
The word changed. The technology did not. For a CIO, that gap is the story.
Superintelligence has had a stable meaning for six decades years and it has never meant a trend. Mathematician I. J. Good, who worked beside Alan Turing at Bletchley Park, defined the “ultraintelligent machine” in 1965 as one “that can far surpass all the intellectual activities of any man however clever,” adding that it would be “the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control.”
Philosopher Nick Bostrom’s 1998 definition was “an intellect that is much smarter than the best human brains in practically every field,” tightened in 2014 to one that “greatly exceeds the cognitive performance of humans in virtually all domains of interest.” Before either, the French scientist Pierre-Simon Laplace’s demon of 1814 was an intellect for which “nothing would be uncertain” [translation].
In every version, the word marks a threshold crossed: a mind beyond ours. The order uses it instead for a direction of travel: systems that “increasingly represent not merely artificial intelligence, but a new era of Super Intelligence.”
Why this matters to you is procurement, not philosophy
Vocabulary is how money moves. Federal solicitations, agency guidance and public sector communications will now carry “SI.” Vendor decks will follow quickly; no sales team keeps saying AI once the customer’s government has stopped. Before long, every platform you evaluate may describe itself as Super Intelligence and the term will convey very little information at all.
Two provisions deserve a calendar entry. Section 2(b) leaves existing regulations, contracts and grants untouched, so your current agreements still say AI and still rely on the 2020 definition.
Section 3(b) gives the Assistant to the President for Science and Technology 60 days, until late November, to propose legislative language for a federal definition of Super Intelligence that “reflects the capabilities described in section 1,” including an assessment of whether it should “modify, expand upon, or otherwise supersede” the statutory definition of AI.
If that definition is written around capability rather than mechanism, it could determine what falls inside existing programs, disclosures and procurement rules. It is the one document in this story to read when it lands.
What I would do this quarter
Write your own definition and put it in every contract. Not the government’s definition and not the vendor’s.
Three tiers do the work: systems that predict, systems that recommend to a human who decides and systems that act. The third tier is where the governance, the audit trail and the off switch belong.
Use those distinctions in board papers, procurement templates and vendor questionnaires and make suppliers map their products to your tiers, not to a label.
Hold claims to numbers. A system described as Super Intelligence should be able to show the benchmark, the methodology and who paid for the test. My company commissions benchmarks and says so on the slide; demand the same of everyone.
Frontier, agentic and now SI are vocabulary words. Latency, accuracy of your data, cost per outcome and failure modes under load are facts.
Build in governance before the name forces the question
Research my company commissioned in September 2026 — with Thinker Media, querying 600 executives across three economies — found that 65% say sovereignty matters more than a year ago and 24% build governance in by default. The gap between those two numbers is the exposure.
A system the government calls Super Intelligence is one your board will ask about by name. Know which rows, which columns and which purposes each system can reach, as well as where the data lives, who holds the keys and who can switch it off without calling the vendor.
Those are engineering decisions and they are cheaper before the headline than after it.
Separate the ambition from the inventory. The order’s purpose section is a statement of national ambition and the ambition is real: Frontier systems do more than imitate tasks. But a CIO manages an inventory.
Most of what you run is tier one and two and it will keep working whatever it is called. Spend the new attention on the tier-three systems where a wrong action costs money and leave the vocabulary to the press office.
Good’s caveat is 61 years old and unimproved: The machine must be “docile enough to tell us how to keep it under control.” Washington has renamed the machine by order. Control was never in the name. It is in the design and the design is still yours.
Michael Gale is Chief Marketing Officer of EnterpriseDB and the author of the Wall Street Journal bestseller The Digital Helix: Transforming Your Organization’s DNA to Thrive in the Digital Age.
