The US executive branch will use 'Super Intelligence' in place of 'artificial intelligence' in future official communications under an order signed by President Donald Trump, The Verge reports. The account describes a terminology change affecting policy websites, policy documents, and press releases.
According to the September 29 report by Lauren Feiner, agencies will not have to revise past regulations or documents. That limit matters: a change to future wording is not the same as rewriting every existing reference or changing what a particular model can do.
A broader label replaces a narrower technical idea
The Verge describes superintelligence as a term used for particularly powerful versions of AI, while reporting that the administration wants to use it in place of the broader statutory definition. The proposed wording therefore collapses a distinction between AI in general and an especially advanced category of system.
Trump presented the change at an event launching America.gov, saying he did not want to use the word 'artificial.' The report connects those remarks with a rebrand he had discussed at the United Nations General Assembly the previous week.
These statements describe how the administration wants to present the technology. They do not establish that today's systems have crossed a technical threshold into superintelligence. Readers assessing a model still need evidence about the tasks it can perform, its limitations, and how that performance was measured.
Self-regulation remains part of the message
The report says Trump signed the order after a White House luncheon with technology executives and government officials. It describes him appearing afterward with Nvidia chief executive Jensen Huang and Elon Musk.
In the quoted remarks, Trump points to 'tremendous self-regulation' and refers to existing institutions including the Department of Justice and FBI. The article does not set out a new technical testing protocol or a new model-performance requirement arising from the terminology change.
That emphasis connects with the separate voluntary AI safety accord and its proposed audits. Franklin's coverage of that agreement distinguishes a company's commitment to internal controls and outside evaluation from evidence that the promised checks are already operating. The naming order and the voluntary accord are different measures; neither should be treated as proof that an individual system has passed a completed safety assessment.
Data centers and local concerns
Trump also addressed the communities where companies seek to build large data centers. The Verge reports that he said firms would work to make communities happy by providing resources and economic support, and predicted that data centers would become popular.
Those comments express an intended approach to local opposition. The supplied report does not specify a funding amount, a project-level agreement, or an implementation schedule. It also does not demonstrate how a change in terminology would resolve concerns around an actual facility.
For readers following federal AI policy, the supported distinction is between the reported communications instruction, the political explanation offered for it, and the separate question of oversight in practice. This article summarizes The Verge's account; it does not provide a clause-by-clause analysis of the underlying order or legal guidance on compliance.
The next useful evidence would be the language agencies actually publish and the concrete oversight measures that accompany it. A stronger label can alter public messaging immediately, but model capability and safety remain questions that require their own evidence.