On September 29, 2026, President Donald Trump signed Executive Order 14434, formally directing federal departments and agencies to stop using the term artificial intelligence and replace it with super intelligence—or SI—across official correspondence, websites, reports, and non-statutory policy documents. The directive states plainly that the executive branch "will not acknowledge" the older phrase in applicable settings, turning what began a week earlier as an off-the-cuff remark at the United Nations General Assembly into binding bureaucratic protocol. Yet beneath the theater of presidential branding lies a high-stakes legal and regulatory maneuver that could reshape how American technology companies are governed, audited, and funded.
Most coverage of the order has treated it as pure political vanity—another instance of a president slapping a superlative onto an existing concept because he dislikes the word "artificial." That reading misses the machinery moving underneath. While the order currently borrows the exact statutory definition of artificial intelligence codified in federal law, Section 3 sets a 60-day clock for the White House science adviser to draft a brand-new legislative definition of Super Intelligence. Depending on how that definition is written—and whether Congress adopts it—entire categories of software, hardware, and frontier research could either be swept into federal oversight or quietly exempted from it.
From UN Stagecraft to Bureaucratic Mandate
The campaign to scrub "artificial" from the federal lexicon did not emerge from a computer science symposium. It started with a gut reaction to a word.
Addressing the United Nations General Assembly on September 22, 2026, Trump told world leaders that the phrase "artificial intelligence" bothered him because it sounded fake. "I don't like anything that's artificial," he told the hall, arguing that the technology was "pure genius" and declaring that American documents would switch to "Super Intelligence". Within 24 hours, a senior State Department official instructed American diplomats to purge "AI" from remarks and press materials. White House Office of Science and Technology Policy Director Michael Kratsios adopted the new acronym on the UN stage the very next day.
Seven days later, the rhetoric hardened into Executive Order 14434, titled Inaugurating the Era of Super Intelligence. The implementation across Washington was immediate and sweeping:
- Defense and Intelligence: The Pentagon renamed its Chief Digital and Artificial Intelligence Office the Super Intelligence Office. Director of National Intelligence Jay Clayton was tapped to chair a new government-wide Super Intelligence Force.
- Law Enforcement: The Department of Justice launched a Super Intelligence Task Force in early November to coordinate civil and criminal enforcement involving algorithmic fraud and attacks on critical infrastructure.
- Science and Standards: The National Institutes of Health began phasing "SI" into its public portals and grant notices, while the Commerce Department stood up a Center for Advancing Innovation and Standards for Super Intelligence.
By October 8, Trump escalated the rhetorical pressure on Truth Social, warning that the White House viewed anyone still using the old phrase as "THE ENEMY". No formal penalties were attached to the post. None were needed. In Washington, procurement officers and grant applicants learn quickly which vocabulary opens doors.
The Technical Friction Behind a Two-Word Swap
Ask a machine learning researcher in Cambridge or San Francisco what "superintelligence" means, and you will get a very different answer than the one printed on White House stationery.
For more than two decades, computer scientists, philosophers, and safety researchers have maintained a strict taxonomy to separate what exists today from what might exist decades from now. Collapsing those boundaries creates immediate conceptual confusion.
| Term | Industry and Academic Meaning | Federal Meaning Under EO 14434 |
|---|---|---|
| Artificial Intelligence (AI) | Software systems built to perform specific tasks requiring human-like cognition, such as text generation, image recognition, or data analysis. | Phased out of non-statutory executive branch communications; "not acknowledged" in official contexts. |
| Artificial General Intelligence (AGI) | A hypothetical system capable of matching human cognitive flexibility across virtually any intellectual domain. | Not separately distinguished in the executive order's terminology shift. |
| Superintelligence (ASI) | A theoretical future system that vastly surpasses the brightest human minds in every field, from scientific creativity to social reasoning. | Applied broadly to all existing technologies currently defined as AI under 15 U.S.C. § 9401(3). |
That distinction matters. Today's frontier models—even those scoring above 96 percent on graduate-level science benchmarks—remain narrow statistical engines. They predict tokens, optimize weights, and still hallucinate basic facts when pushed outside their training distributions. Calling a customer-service chatbot or a logistics routing algorithm "super intelligence" stretches the English language past its breaking point.
Critics inside the research community argue that the rebrand blurs the exact risk thresholds lawmakers need to track. When every automated spreadsheet tool is labeled "super intelligence," policymakers lose the vocabulary needed to describe genuine frontier breakthroughs that actually warrant heightened scrutiny.
Proponents of the move counter that the 1955 term "artificial" was always a relic of early academic modesty. When John McCarthy and his colleagues coined the phrase for the Dartmouth workshop seventy years ago, computers struggled to play basic checkers. In 2026, models write production code, accelerate drug discovery, and run targeting platforms for the Pentagon. Calling those outputs "artificial," supporters argue, undersells American engineering at the exact moment the United States is locked in a fierce technological race with China.
Why Silicon Valley Played Along So Quickly
If tech executives were privately baffled by the sudden vocabulary mandate, few showed it in public. Within days of the executive order, major industry players began scrubbing their own branding to match the Oval Office.
Elon Musk announced he would rename SpaceX's AI unit to SpaceXSI. Weeks later, as Anthropic prepared for its initial public offering in November, the company updated its amended S-1 regulatory filing to use "artificial intelligence" and "super intelligence" interchangeably—a sharp departure from earlier drafts that barely mentioned the latter phrase.
Silicon Valley's compliance was not driven by sudden agreement on semantics. It was transactional.
On the exact same day Trump signed Executive Order 14434, six major technology leaders stood beside him to sign the White House Accord on Super Intelligence. That 300-word agreement is a voluntary pledge built around four lightweight governance layers—internal monitoring, a designated safety owner, external audits, and board oversight—with zero statutory penalties or federal enforcement teeth.
In exchange for adopting the administration's preferred branding and signing a non-binding pledge, the industry secured powerful political backing against stricter regulation. Vice President JD Vance framed the accord as proof that developers should police themselves rather than ask Washington for new rules. Meanwhile, Senate efforts to establish a statutory AI Safety Board stalled, and the White House maintained pressure against state-level technology laws. For executives staring down billions of dollars in data-center permitting and federal defense contracts, swapping an "A" for an "S" was the cheapest lobbying victory in town.
The November 28 Deadline Nobody Is Watching
Look closely at the text of Executive Order 14434, and the real policy trap appears in Section 3.
Right now, the rename is legally cosmetic. The order explicitly ties "Super Intelligence" to 15 U.S.C. § 9401(3)—the existing statutory definition of artificial intelligence enacted by Congress in the National AI Initiative Act of 2020. Existing contracts, grants, and historical regulations remain untouched.
That freeze is temporary.
Section 3 orders the Assistant to the President for Science and Technology to deliver proposed legislative language for a standalone federal definition of "Super Intelligence" within 60 days—putting the deadline at November 28, 2026. The science adviser must also recommend whether that new definition should modify, expand, or completely replace the current statutory text.
This is where semantics turn into hard law. Consider what happens if the White House proposes a narrower definition tailored to frontier capabilities—and a friendly Congress writes it into the U.S. Code.
- Regulatory Carve-Outs: Hundreds of federal procurement rules, export controls, and civil-rights guidelines rely on the statutory definition in Section 9401(3). Narrowing the definition under the banner of "Super Intelligence" could instantly exempt conventional enterprise software from existing federal compliance mandates.
- Export and Chip Controls: Commerce Department rules governing semiconductor shipments to the Middle East and Asia depend on precise technical thresholds. Redefining the core term alters which compute clusters trigger national-security reviews.
- Procurement Confusion: Federal contractors are already navigating a split screen. Agencies will not acknowledge "artificial intelligence" in new bids or proposals, yet contractors must still cite the old statutory phrase when referencing existing laws and active grants.
The Global Friction of a Unilateral Dictionary
Technology standards do not stop at the water's edge, and Washington's new dictionary collides directly with the rest of the industrialized world.
The European Union's AI Act, the United Nations' global advisory bodies, and bilateral research frameworks with allies in Tokyo, London, and Seoul are all anchored in the term "artificial intelligence." When State Department cables instruct diplomats to scrub "AI" from international communiqués, treaty negotiators face an immediate translation problem. Foreign ministries cannot simply find-and-replace terminology embedded in ratified trade agreements and safety protocols without reopening settled text.
Even inside the federal government, the transition exposes deep institutional seams. The National Institutes of Health warned researchers in October that legacy databases and PubMed citations would continue displaying both terms indefinitely while public-facing web pages undergo revision. Law firms advising defense and healthcare contractors have begun issuing formal memos warning clients to mirror "Super Intelligence" in new agency pitches while preserving "artificial intelligence" in statutory compliance filings to avoid legal ambiguity.
What Happens When the Label Outruns the Reality
History offers plenty of warnings about governments that rewrite technical vocabulary to project strength. Words can rally budgets, align agencies, and flatter national pride. They cannot alter the underlying math of a neural network.
By ordering the executive branch to speak as if superintelligence has already arrived, the White House has handed Silicon Valley a powerful marketing gift while handing regulators a blunter instrument. Until Congress acts on the legislative proposal due later this month, every federal agency is operating under a linguistic split personality—enforcing laws written for artificial intelligence while issuing press releases celebrating super intelligence. Whether that gap closes through deregulation or collapses under the weight of its own hype will depend entirely on the fine print delivered to the Oval Office on November 28.
