Trump signs an order renaming artificial intelligence 'super intelligence' across the US government

What the order says

On 29 September 2026, President Trump signed an executive order titled "Inaugurating The Era Of Super Intelligence." Its central instruction is unusual for a technology policy: it is about vocabulary. The order directs every executive branch department and agency to use the terms "Super Intelligence" and "SI" in place of "Artificial Intelligence" and "AI" in official correspondence, communications, websites, reports and policy documents. It goes further than a style preference, stating that agencies "will not acknowledge the usage of 'Artificial Intelligence' and 'AI' in any applicable setting."

The order does carry one forward-looking task. It gives the Assistant to the President for Science and Technology 60 days to submit proposed legislative language establishing a Federal definition of "Super Intelligence," including an assessment of whether that definition should modify existing statutory definitions of AI. It also carves out an exception: departments do not have to rewrite previously issued regulations, presidential actions, contracts, grants or other historical documents.

The stated reasoning, in the order's own words, is that modern systems "do much more than imitate or automate discrete aspects of human intelligence" and now represent "not merely artificial intelligence, but a new era of Super Intelligence."

A new name, not a new technology

The most important thing to understand about the order is what it does not do. It does not change a single capability of any system the government uses. The models, tools and services that agencies ran on 28 September work identically under their new label. Nothing was trained, upgraded or restricted. What changed is the word.

That distinction matters because a rename like this is easy to read as a signal that the technology has crossed some threshold. It has not. An executive order can direct what the government calls a technology; it cannot legislate the technology into a new category of capability.

The word already means something

There is a further complication, and it is not a small one. "Superintelligence" is not a blank term waiting to be assigned. It has a settled meaning in computer science and AI safety, and it is a demanding one. The philosopher Nick Bostrom's 2014 book "Superintelligence: Paths, Dangers, Strategies" popularized the standard definition: an intellect that greatly exceeds human cognitive performance in virtually every domain. Researchers use it to describe a hypothetical future system far beyond today's models, the thing that comes after human-level AI, not the chatbots and copilots in use now.

By that established meaning, no current system is a superintelligence, and most researchers do not believe one exists yet. Applying the label to all of today's AI, by government fiat, collides with how the field uses the word. It risks an odd situation in which federal documents describe ordinary machine-learning tools with a term the research community reserves for a technology that has not arrived. A word that once flagged the far horizon of the field would, in official US usage, come to mean whatever AI already is.

The safety accord signed the same day

The rename did not happen in isolation. Hours before, at the White House, the leaders of six major technology companies signed a separate frontier-model safety accord: Google's Sundar Pichai, Anthropic's Dario Amodei, Meta's Mark Zuckerberg, OpenAI president Greg Brockman, xAI's Elon Musk and NVIDIA's Jensen Huang. Under the accord, companies agreed to four layers of controls: internal monitoring of model capabilities and alignment, internal oversight teams, independent external audits, and independent committees to review those audits. Musk described it as agreeing to "joint monitoring, board special committees, just generally grading each other's homework."

The accord is voluntary, and the administration framed it as an alternative to regulation. Vice President Vance put the philosophy plainly: "The solution to some of the AI risks is for you guys to take the risk seriously, not to come to the government for a regulatory regime." The pairing is telling. On the same day the government adopted the most maximal possible name for the technology, its actual safety approach leaned on companies policing themselves. For a sense of how seriously the labs themselves treat these risks, Anthropic's own IPO prospectus warned that advanced AI could pose catastrophic or existential dangers, and NVIDIA has shipped tooling to contain rogue AI agents.

Our take

Names shape how people think, so a government insisting on "super intelligence" is not nothing. But it is worth being precise about what happened. This is a branding decision with the force of an executive order, not a technical event. The capabilities of the systems are unchanged, and the word chosen already describes something more advanced than anything that exists. There is a real policy thread inside it, the 60-day push to define "super intelligence" in law, which could matter if it reshapes how statutes treat the technology. But the headline act is linguistic. The most useful response for anyone following AI is to keep measuring the systems by what they can actually do, and to treat the label, whichever one the government prefers, as marketing until the capabilities catch up to it.