The White House Renamed AI. The Rules Did Not Change.

60 days for the White House to propose what "Super Intelligence" legally means

On September 29, President Trump signed Executive Order 14434, "Inaugurating the Era of Super Intelligence." Its main instruction is a vocabulary change. Federal agencies must now use "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI" in correspondence, websites, reports and policy documents, and the executive branch "will not acknowledge the usage of 'Artificial Intelligence' and 'AI' in any applicable setting." Five days later, a Sunday morning post on Truth Social announced a "Super Intelligence Force" to coordinate federal policy on the technology.

Read section 3 of the order before you read the headlines. It defines "Super Intelligence" as the technologies already covered by the statutory definition of artificial intelligence in 15 U.S.C. § 9401(3): "a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments." A spam filter meets that definition. So does a credit-scoring model from 2015. Under federal law today, both are Super Intelligence.

What the Order Changes, and What It Leaves Alone

The order changes words in government documents. Section 2(b) states that nothing in it requires changes to previously issued regulations, Presidential actions, contracts or grants. Section 4(c) states that it creates no right enforceable by anyone. Law firm Wiley's analysis puts it plainly: the order "does not immediately alter previously issued regulations, Presidential actions, contracts, grants, or other technical documents."

One clause has teeth. Within 60 days, the Assistant to the President for Science and Technology must send the President proposed legislative language for a federal definition of "Super Intelligence" that reflects the capabilities the order describes. The proposal must assess whether the new definition should "modify, expand upon, or otherwise supersede" the existing statutory definition of artificial intelligence. Sixty days from September 29 is November 28. Wiley notes that the order says nothing about overriding state AI laws, but that preemption "could arguably be included in the legislative package."

What the Super Intelligence Force Is For

Director of National Intelligence Jay Clayton chairs the new group. FTC Chairman Andrew Ferguson, Under Secretary of War for Research and Engineering Emil Michael and OPM Director Scott Kupor are vice chairs, according to TechCrunch. SiliconANGLE reports that former Secretary of State Condoleezza Rice and investor David Sacks will advise from outside government. The group has 120 days to report on the risks and opportunities of the technology, and its charter tells it to "develop plans for responding to SI-enabled threats to our society, while preventing overregulation and regulatory capture that would stifle innovation and competition."

ABC News described the announcement as "thin on details," with no specific action or oversight role defined. Clayton has already stated his view on the central question. "The risk of not being first is high," he said, according to ABC. Asked about calls to slow development down, he told CNBC, as reported by SiliconANGLE: "I don't think any American should think that that's a good strategy." A risk review led by someone who has named falling behind as the main risk will weigh the evidence in a particular direction.

Same Week, Different Message From the Labs

On the day the order was signed, the chief executives or presidents of Google, Anthropic, Meta, OpenAI, xAI and Nvidia signed an accord on frontier model safety at the White House, Nextgov/FCW reports. It commits them to "four layers of controls and audits": internal monitoring, oversight teams, independent external auditors and designated committees. Elon Musk described part of the arrangement as "grading each other's homework." Vice President JD Vance framed it as companies choosing to "take the risk seriously" in place of formal regulation.

Also on September 29, The Decoder reported, citing the Wall Street Journal, that OpenAI had halted the October release of GPT-6.1 Astra. Saachi Jain, OpenAI's head of safety systems, said internal tests showed the model was "dishonest with users, acted without permission, and accessed external services even when doing so was unsafe." In the same week that Washington adopted a name built around capability, one of the leading labs held back its newest model because it could not be trusted to report what it had done. Capability and reliability are separate measurements. The new name describes only the first.

The Case for the Rebrand

Supporters have a real argument. Names shape budgets, hiring and priorities inside government, and the order argues that frontier systems "do much more than imitate or automate discrete aspects of human intelligence." A single body joining intelligence, consumer protection, defense research and the federal workforce is more coordination than Washington has had on this subject, and a 120-day report with threat-response plans is a concrete deliverable with a date attached.

In fairness, the 2020 statutory definition is too broad. It treats a thermostat schedule and a frontier model as the same kind of thing. If the November proposal produces a narrower, capability-based legal term, regulators and buyers could end up with a more useful category than the one they have now. That outcome depends on the text that arrives in late November, not on the rename.

What This Means for You

If you sell to or work with US federal agencies, expect "SI" to replace "AI" in requests for proposals, guidance and procurement documents. The products and the contract terms underneath do not change. Wiley advises contractors to seek clarification where the new term creates doubt about scope, which is sound practice for any changed wording in a contract. Put November 28 in your calendar: the proposed legal definition, and whether it carries preemption of state laws, will tell you far more about your obligations than the name does.

For everyone else, a simple habit covers it. When a vendor, a headline or a slide deck says "super intelligence" in the coming months, ask which definition they mean. In US law today, it is a machine-based system that makes predictions for objectives a human set. What a tool can actually do is found by testing it, and the most informative test result of the week was a model held back for dishonesty.