Michael JohnsonThe September 29 order changes the official language of the executive branch; it does not confirm the arrival of superhuman intelligence. A separate voluntary agreement among technology companies cannot be fully assessed without its text.
Signed by Donald Trump on September 29, 2026, the order directs executive agencies to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official correspondence, public communications, websites, reports, and other documents not specified by law. The existing statutory definition of artificial intelligence remains in force for purposes of implementing the order.
This is not a nationwide renaming of the technology, nor evidence that current models have achieved superhuman capabilities. The order changes the vocabulary used by the federal executive branch, while the development of a new definition has merely been commissioned.
What the order actually requires
The requirement applies to executive departments and agencies and takes effect “to the maximum extent permitted by law.” The order does not require previously issued documents to be rewritten, including regulations, presidential documents, contracts, grants, and historical materials.
For purposes of the directive, “Super Intelligence” and “SI” cover the same technologies and systems as the current federal definition of AI. The order calls for a draft of a new definition: within 60 days, the President’s science and technology adviser must submit a draft statutory definition of the new term and proposals for possible changes to the law. This is an assignment to prepare proposals, not a legislative change that has already been enacted.
The White House justifies the renaming by arguing that, in the administration’s assessment, advanced systems no longer merely imitate or automate individual human abilities, but expand people’s capabilities. This is the administration’s position, as set out in the White House fact sheet, not an independent assessment of what the systems can do.
A new name is not proof of new capabilities
The order’s immediate effect is to change federal agencies’ documents and public communications. The new term may appear where “AI” was previously used, while older documents and references to current law retain the existing terminology.
A name change alone says nothing about the accuracy, safety, or capability level of any particular model. The order directs agencies to use the new terms and calls for proposals for a new federal definition, but it does not establish that all covered systems have become superhuman.
The companies’ agreement is a separate matter
At a White House meeting, technology company leaders discussed internal procedures and external reviews. However, according to the Associated Press report on the voluntary agreement, its text was unavailable at the time of publication, and details of the commitments had not been disclosed. The participants’ statements therefore do not establish the companies’ precise obligations, review criteria, auditors’ independence, or consequences for failing to meet their promises.
This is an initiative separate from the order: the order concerns terminology used by federal agencies, while the agreement concerns a stated voluntary commitment by businesses. Without access to the text, it is impossible to assess whether the promised reviews amount to a verifiable oversight system.
Political interpretation
In an analysis published October 1, Axios interprets the new name as an attempt to change public perceptions of AI amid debates about the technology’s risks and data center construction. That is the publication’s interpretation, not a motive explicitly established by the order.
More precisely, the administration has directed federal agencies to use a new term in some official materials while retaining the current statutory definition as the basis for implementing the order. It remains unclear what will change after a draft of a new definition is prepared, or whether companies’ commitments will become publicly verifiable.