On 29 September 2026 the White House issued an executive order titled Inaugurating the Era of Super Intelligence. It directs federal departments and agencies to use "Super Intelligence" and "SI" in place of "artificial intelligence" and "AI" in official correspondence, public communications, policy documents and other non-statutory materials.
If you run validation or quality in a regulated company, the useful question is narrow: does anything you are responsible for change? Almost certainly not. Here is what the order says, what it deliberately does not touch, and the small number of practical consequences.
Written 30 September 2026, one day after signing. A federal definition of SI has been commissioned but not yet published, so treat this as a developing picture.
What the order actually does
- It changes vocabulary inside the executive branch. Agencies are to use "Super Intelligence" and "SI" in correspondence, public communications, websites, reports and policy documents, to the maximum extent permitted by law.
- It keeps the meaning identical. This is the part most coverage skipped. Section 3(a) defines SI as the technologies and systems already encompassed by "artificial intelligence" as defined in 15 U.S.C. §9401(3). Same scope, new label.
- It does not rewrite the past. Section 2(b) states that nothing in the relevant section requires altering previously issued regulations, Presidential actions, contracts, grants or historical documents.
- It commissions a definition. The Assistant to the President for Science and Technology is to propose a federal definition of SI and identify any further executive action needed.
A separate industry accord was signed the same day by a group of technology chief executives. It is voluntary — the President described it as "morally binding" rather than legally enforceable — and commits signatories to internal controls and multiple layers of audits and reviews. We look at that separately in the White House SI accord.
What it does not do
An executive order of this kind cannot amend statute or regulation, and this one explicitly does not try to. So:
- 21 CFR Part 11 is unchanged, word for word. So are 21 CFR 210/211 and Part 820.
- EU GMP is entirely outside its reach — Annex 11, the draft Annex 22, the EU AI Act and MHRA guidance all still say artificial intelligence. See US says SI, Europe says AI.
- GAMP 5 and ISPE guidance are unaffected. They are industry documents, not federal ones.
- Your validation obligations are identical. Nothing about intended use, risk assessment, testing, human oversight or change control moves because a word changed. See what has not changed.
As of 30 September 2026 the FDA had not published guidance using the new terminology. Existing FDA documents on AI in drug and device contexts still read "artificial intelligence", and the order does not require them to be reissued.
Where it will actually reach you
Three places, over time, and none of them urgent:
- New FDA communications. Future guidance, web pages and correspondence may use SI. When they do, read them as meaning exactly what AI meant — the order says so explicitly.
- Inspection conversation. An investigator may use either word. Neither changes what they are entitled to see.
- Vendor marketing. This is the fastest-moving one, and the least useful. Expect "SI-powered" on slides within the week — covered in SI washing.
What you should not do is launch a documentation project. A find-and-replace across your SOPs creates a change-control burden, a training obligation and a version history full of cosmetic revisions, in exchange for nothing. See do you need to rename AI in your SOPs.
The sensible response
Add a single line to your AI governance or validation procedure noting that US federal materials may use "Super Intelligence (SI)" for what the document calls artificial intelligence, and that the terms are equivalent for the purposes of that procedure. One sentence, one controlled change, no retraining.
Then carry on. The work that makes an AI system defensible in a regulated environment is unchanged: a written intended use, risk proportionate to consequence, honest test data, a qualified person accountable for the output, and monitoring after go-live. That is the subject of GxP AI validation, and none of it turns on what the technology is called this month.
Primary source: the White House fact sheet on the order. Wider context in GxP AI.
Where to go next
Explore GxP Copilot for AI-native validation, TraceDraft for source-traceable clinical documentation, or book a demo to see either on your own data.