During an address to the United Nations General Assembly on Tuesday, Trump announced that the U.S. government would begin referring to artificial intelligence as “Super Intelligence,” or SI, arguing that the word “artificial” no longer accurately describes the technology.
“From this point forward, all of United States documents, and hopefully the world's, will be changed to use the much more accurate term ‘super’ as opposed to artificial,” Trump told delegates. “Welcome to the new world of super intelligence — S.I.”
Trump had already been floating the new terminology before his UN appearanceIn a Truth Social post a day earlier, he referred to the technology as “AI or, SI (SUPER INTELLIGENCE)” while arguing that the U.S. should continue accelerating development rather than slowing it over concerns about advanced AI.
The name change might sound mostly cosmetic. But there’s one big problem: “superintelligence” already means something in AI. And it definitely isn't ChatGPT.
AI and superintelligence aren't the same thing
Artificial intelligence is an umbrella term covering a huge range of technologies, including the generative AI tools millions of people now use every day.
Superintelligence is something considerably more specific. AI researchers generally separate machine intelligence into a few broad categories.
- Artificial narrow intelligence (ANI) describes AI designed to perform particular tasks. That includes much of the technology available today, from image recognition systems to large language models and generative AI assistants.
- Artificial general intelligence (AGI) generally refers to a hypothetical AI capable of matching human-level intellectual abilities across a broad range of tasks, rather than being limited to particular areas.
- Artificial superintelligence (ASI). This is the theoretical step beyond AGI: a system whose abilities exceed those of humans across essentially every intellectual domain. Nobody has demonstrated that technology yet.
IBM describes artificial superintelligence as a “hypothetical” AI system with intellectual abilities beyond those of humans. Google similarly describes ASI as a theoretical form of AI that would surpass human intelligence.
That means changing the label on today's AI doesn't change what the technology can actually do. ChatGPT doesn't suddenly become superintelligent because a government document calls it SI. Neither do Claude, Gemini or any of the other AI assistants available today.
Why the phrasing matters
Things might get confusing really quickly. Just imagine seeing a government report describing the risks of “super intelligence.” Is it talking about ChatGPT and Gemini as they exist today? More autonomous AI agents capable of taking actions on someone's behalf? Or a hypothetical future system more intelligent than any human?
Under the terminology Trump proposed Tuesday, those very different technologies could potentially end up sharing the same label.
There’s also a consumer perception problem. Today's most advanced AI models are extraordinarily capable, but they're also far from infallible. Chatbots can still hallucinate information, misunderstand instructions and confidently produce incorrect answers. AI agents can also make mistakes while using software and carrying out multistep tasks.
Calling those systems “super intelligence” could give users a very different impression of what they're capable of.
The announcement is also about the AI race
The president used the speech to reiterate his opposition to international restrictions that he believes could slow U.S. AI development and framed leadership in advanced AI as a competition between the United States and China.
That follows comments Trump made a day earlier in which he said the U.S. would be careful about AI and could use the Justice Department and other law-enforcement agencies to intervene if necessary, while insisting that his administration would continue encouraging its development.
The remarks come as other world leaders are pushing for more international coordination around increasingly capable AI.
A group of countries and institutions called this week for AI to remain under human direction, supervision and control and for governments to develop common standards around the technology.
That makes the debate about “SI” bigger than a new acronym because the disagreement is also about how quickly powerful AI should be developed, how it should be regulated and whether international rules should play a role.
Is AI actually becoming superintelligent?
There's no question that AI models are getting more capable. The latest systems can write software, analyze images, operate computers, use tools and complete increasingly complicated tasks with less human involvement.
But capability isn't the same thing as superintelligence.
Even artificial general intelligence (AGI) remains a disputed threshold with no universally accepted test for determining when it has been reached. Superintelligence goes significantly further by describing systems that surpass human intellectual ability broadly rather than simply outperforming people at particular tasks.
Today's AI can already outperform humans in certain narrow areas. That doesn't make it ASI, which means the biggest question is whether Trump's announcement actually changes the terminology used across the federal government.
Trump said U.S. documents would switch from “artificial intelligence” to “super intelligence,” but at the time of publication, the announcement itself doesn't change the terminology used by the rest of the technology industry.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom's Guide on YouTube and follow us on TikTok.
More from Tom's Guide
