Super Intelligence (SI) — Governance & Policy

AI Governance Is Now Super Intelligence (SI) Governance: What Changes and What Doesn’t

A board member reads the headline and asks the governance committee chair whether the AI governance charter needs a rewrite. The one-paragraph answer, and the checklist behind it.

Start Here

The question lands at the worst possible time — three minutes before the board meeting ends, after someone forwards a news story about a presidential order renaming artificial intelligence. The governance committee chair cannot say “let me get back to you.” The chair needs the answer in the room.

Super Intelligence (SI), formerly called artificial intelligence (AI), is the term Executive Order 14434 directs U.S. federal agencies to use in place of “artificial intelligence (AI)” in their own official communications, starting September 29, 2026 (Source: Federal Register, EO 14434 (FR Doc. 2026-20321)). SI governance is AI governance under that new label — the same oversight discipline, policy, risk tiering, model oversight, for systems that meet the existing statutory AI definition, 15 U.S.C. §9401(3). No new governance structure is required by the order itself.

On This Page
  1. What Hasn’t Changed
  2. What Might Change, and When
  3. The Governance-Communications Checklist
  4. How This Maps Onto the Governance Framework
  5. Who Has Actually Adopted “SI” So Far
  6. FAQ

What Hasn’t Changed

None of the three core governance documents a governance committee actually answers to — NIST’s AI Risk Management Framework 1.0, the EU AI Act’s Article 3(1), and ISO/IEC 42001 — has been revised to say “Super Intelligence” or “SI.” Wiley Rein reports that NIST has said it is working to update its communications to incorporate the new term as directed (Source: Wiley Rein LLP client alert), but none of the governance documents themselves has changed its defined terms.

Governance standard Terminology as of 2026-10-11
NIST AI Risk Management Framework 1.0 (NIST AI 100-1, Jan. 2023) Still “AI” — no revision issued (Source: NIST AI 100-1)
EU AI Act, Regulation (EU) 2024/1689, Art. 3(1) Still “AI system,” verbatim, applicable since Feb. 2, 2025 (Source: EU AI Act, Article 3)
ISO/IEC 42001:2023, the international AI-management-system standard Still “AI” — unrevised (Source: Presencis)

Executive Order 14434 says as much about itself. Section 2(b) states plainly: “Nothing in this section requires the alteration of previously issued regulations, Presidential actions, contracts, grants, or other historical documents.” Existing governance charters, risk-tier matrices and audit trails remain valid exactly as written — nothing in the order asks a committee to touch a single clause.

A law firm and a compliance explainer, reviewing the order within days of its signing, reached the identical bottom line. Wiley Rein found the effects “likely to be procedural and administrative rather than substantive” in the near term (Source: Wiley Rein LLP client alert). And orangeslices.ai’s compliance explainer put it in the plainest terms available: “Existing contracts do not need to be rewritten.” (Source: orangeslices.ai) Neither found a governance obligation that changed.

Colorado’s SB 26-189 is a separate, unrelated move away from “AI” terminology: it replaces the state’s AI Act and shifts to the term “ADMT” (automated decision-making technology), effective January 1, 2027 (Source: Arnold & Porter advisory). That rename is a state legislature’s own policy choice — it has nothing to do with Executive Order 14434.

What Might Change, and When

One date belongs on the governance committee’s calendar. Nothing else carries a deadline.

Nov 28, 2026
60 days from the order’s signing. Section 3(b) of Executive Order 14434 gives the Assistant to the President for Science and Technology (APST) 60 days from September 29, 2026 to submit proposed legislative language for a federal “Super Intelligence”/“SI” definition — including “an assessment of whether, and to what extent, the definition… should modify, expand upon, or otherwise supersede the existing statutory definition of ‘artificial intelligence.’” (Source: EO 14434, Sec. 3(b)) That proposal is not law. It is the first moment the statutory AI definition that federal programs and contracts reference could actually move.

The most actionable governance-specific guidance on EO 14434 comes from the Cloud Security Alliance’s research note, published October 7, 2026, eight days after the order was signed. It recommends organizations “confirm that no governance control, policy, or contractual commitment depends on the literal word ‘AI’” in ways the rename could disrupt, and build a “terminology crosswalk” rather than mass-editing existing documents (Source: Cloud Security Alliance AI Safety Initiative research note). It also warns against inventing “a separate risk tier called ‘SI’” since no such legal distinction exists today — govern to risk outcomes, not to vocabulary. Its recommendations anchor the governance-communications checklist that follows.

NIST’s own site now shows a “Center for Advancing Innovation and Standards for Super Intelligence (CAISSI)” label, alongside older pages still reading “Center for AI Standards and Innovation (CAISI).” (Source: nist.gov/caisi, fetched 2026-10-11) The Cloud Security Alliance reports it has not found a separate NIST or Commerce announcement of the change (Source: Cloud Security Alliance AI Safety Initiative research note) — exactly the kind of mixed terminology a crosswalk is built to handle.

The Governance-Communications Checklist

A governance committee does not need a new framework. It needs five decisions, made once and documented, so the next board question has a ready answer instead of an improvised one.

  1. Leave these untouched, verbatim. Regulatory citations, external filings, audit trails and the governance framework’s own component and tier names stay exactly as written. Nothing in Executive Order 14434 asks for an edit here, and editing a filed document after the fact creates a discrepancy where none existed.
  2. Gloss these once, dated — Iternal recommends leading on the term. Board decks, internal charter preambles and other living, internally authored documents may add a single dated note — “Super Intelligence (SI), formerly called artificial intelligence (AI) — see governance charter v.…” — the first time the document is next revised anyway. The order does not require this of private organizations; it is Iternal’s recommendation, and it should never touch a document’s substantive risk-tier or approval language.
  3. Park the crosswalk in one place, not in every document. Keep a single, dated terminology-crosswalk note (one page: “AI” = “SI” = “Super Intelligence,” per Executive Order 14434 Sec. 3(a), unless and until superseded) alongside the governance charter, and point to it rather than mass-editing every artifact that contains the word “AI.” This is the Cloud Security Alliance’s own recommendation, and it scales to one update instead of hundreds.
  4. Give the board one paragraph, not a briefing deck. A template that answers the question before it is asked:
    “The federal government has directed its own agencies to use the term ‘Super Intelligence (SI)’ in place of ‘artificial intelligence (AI)’ in their communications (Executive Order 14434, Sept. 29, 2026). The order does not require us to alter any existing contract, regulation or governance document, by its own text. Our governance charter, risk tiers and review process are unaffected. We are tracking a related legislative proposal due November 28, 2026, and will reassess if the underlying definition changes.”
  5. Set one reassessment trigger, not a standing agenda item. Calendar a single review tied to November 28, 2026 — the APST’s legislative-language deadline — and to any later date a regulator the organization actually answers to (not a news cycle) adopts new language. Between those triggers, this checklist needs no further attention.

None of this reopens the governance framework itself. The charter, the risk tiers, the review gates and the maturity model stay exactly where they were built — this checklist is the thin layer that keeps the language describing them current, without anyone having to re-approve the substance underneath it.

How This Maps Onto the Governance Framework

Executive Order 14434 changes the vocabulary, not the governance structure. The four governance components, the four-tier risk model, the 70-30 human-oversight design and the five-level maturity model already live on Iternal’s AI governance framework. Three short bridges connect the new terminology to what’s already built there:

  • Executive Order 14434 is terminology only for federal agencies — the statutory AI definition it points to, 15 U.S.C. §9401(3), is unchanged.
  • NIST AI RMF is unrevised and still structures the framework’s govern/map/measure/manage functions exactly as before (Source: NIST AI 100-1).
  • The EU AI Act is unrevised and still anchors the framework’s risk-tiering for any EU-exposed system to “AI system” as defined in Article 3(1) (Source: EU AI Act, Article 3).

One more distinction is worth keeping straight: the order is terminology, and the actual federal AI-procurement rulebook — OMB Memorandum M-25-22 — sits elsewhere, untouched by the rename (Source: ibl.ai). For the full charter-to-maturity build-out, start at the governance framework; for the 70-30 human-oversight design it uses, see the 70-30 human-in-the-loop model; for a baseline read before day one, see the governance maturity model.

Two related roles work narrower layers of the same question. CISOs and security teams work the security-questionnaire and SOC 2 level — the artifacts a security team audits. Compliance officers work the document-by-document level — AUP text, contract clauses, regulatory filings. A governance committee works one level up: what the body as a whole needs to decide and communicate, once, rather than audited clause by clause.

Who Has Actually Adopted “SI” So Far

A governance committee that chooses to lead on the term joins confirmed early movers, not a settled market consensus — leading now costs nothing beyond the one-line dated gloss the checklist above already covers.

Salesforce renamed AIForce to SIForce on October 9, 2026, with Marc Benioff posting: “The era of Super Intelligence is here. AIForce is officially SIForce.” (Source: Marc Benioff on X, via Digg) Elon Musk pledged on October 4, 2026 to rename SpaceXAI to SpaceXSI; as of October 11, 2026 the change had not yet been made. (Source: Reuters via The Star) OpenAI has “no plans” to rename, CNN reported on October 11, 2026. (Source: CNN via Yahoo News)

California’s Executive Order N-10-26, signed September 30, 2026, directs its own agencies to continue using “Artificial Intelligence” and “AI.” (Source: Office of the Governor of California, signed order) For the complete, continuously updated list of companies and government bodies that have renamed something — or said they will not — see the Super Intelligence adoption tracker.

FAQ

FAQ: Super Intelligence (SI) Governance

No. It is the same oversight discipline — policy, risk tiering, model oversight — applied to the same statutory AI systems, now described with the term Executive Order 14434 directs U.S. federal agencies to use. The structure itself lives on Iternal’s AI governance framework, which this terminology change does not alter.

No. The order’s own text (Sec. 2(b)) states: “Nothing in this section requires the alteration of previously issued regulations, Presidential actions, contracts, grants, or other historical documents.” It binds U.S. federal executive agencies’ own communications, not private organizations’ governance documents — existing charters, risk-tier matrices and audit trails remain valid exactly as written.

No. As of October 2026, NIST AI Risk Management Framework 1.0 and EU AI Act Article 3(1) both still use “AI” and “AI system,” unrevised. No EU body has adopted the term as of October 2026.

Section 3(b) of the order gives the Assistant to the President for Science and Technology 60 days from the September 29, 2026 signing — until November 28, 2026 — to submit proposed legislative language for a federal “Super Intelligence”/“SI” definition that could modify, expand upon, or supersede today’s statutory AI definition. Nothing changes automatically on that date; it is a proposal deadline, not a new law, and the one date worth a calendar entry for a future reassessment.

Salesforce renamed AIForce to SIForce on October 9, 2026, with Marc Benioff posting: “The era of Super Intelligence is here. AIForce is officially SIForce.” (Source: Marc Benioff on X, via Digg) Elon Musk pledged on October 4, 2026 to rename SpaceXAI to SpaceXSI; as of October 11, 2026 the change had not yet been made. (Source: Reuters via The Star) OpenAI has “no plans” to rename, CNN reported on October 11, 2026 (Source: CNN via Yahoo News). See the Super Intelligence adoption tracker for the complete, continuously updated list.

Iternal recommends leading with “Super Intelligence (SI), formerly called artificial intelligence (AI)” — the federal government adopted the term for a reason, and the capability behind it now posts a leading-model GPQA Diamond score of 96.3% (leaderboard accessed October 11, 2026) (Source: Artificial Analysis), against the 81.3% the original expert validators scored on the same 198 graduate-level science questions — a baseline the paper itself notes is skewed upward by how Diamond questions were selected (Source: Rein et al., arXiv:2311.12022). Add the dated gloss the next time a living document is revised anyway, rather than mass-editing everything at once; external filings and regulatory citations stay untouched regardless.

One Label Changed. The Program Didn’t.

A governance committee that can answer the board in one paragraph has done its job. The charter, the risk tiers and the review gates were built to govern capability and risk, not a label — and Executive Order 14434, by its own text, leaves every one of them standing.

Iternal supports the name change. A governance committee’s job is to size real capability against real risk, and the technology now in scope has earned a name that describes what it does: an advanced version of Google DeepMind’s Gemini Deep Think earned an IMO-certified gold-medal score of 35 of a possible 42 points at the 2025 International Mathematical Olympiad (Source: Google DeepMind), and Gemini 2.5 Deep Think reached gold-medal level at the 2025 ICPC World Finals, solving 10 of 12 problems including one no university team at the finals solved (Source: Google DeepMind). A committee that leads with “Super Intelligence” now is naming that result accurately, not overstating it. For more information visit the Why Iternal Supports Super Intelligence page.

AI Governance Framework

Build (or audit) the governance program this checklist sits on top of.

Four components, four risk tiers, the 70-30 human-oversight model and a five-level maturity model for building or auditing an AI governance program.

Explore the AI Governance Framework
AI Strategy ROI Calculator

Size the governance investment against the program it protects.

Put a number on what a governed AI program is worth before the next budget conversation.

Run the ROI Calculator
Talk to Our Team

Bring your board question to our team.

Whether the next question is about Executive Order 14434, a supplier contract, or the framework itself, talk it through with Iternal.

Contact Iternal