Trump Has Renamed AI “Super Intelligence.” But Does SI Already Mean Something Else?

On September 29, 2026, President Donald Trump signed Executive Order 14434. The order directs U.S. executive departments and agencies, to the maximum extent permitted by law, to use “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official correspondence, websites, policy documents and other non-statutory material. Existing regulations, contracts and historical documents do not have to be rewritten. The White House

The White House argues that “artificial” no longer captures what modern systems can do, and that “Super Intelligence” better reflects their growing capabilities and economic potential. The White House That creates an unusual terminology problem. Superintelligence already meant something else.

The old meaning of superintelligence

Long before the September order, AI researchers used “superintelligence” to describe a hypothetical level of machine intelligence that substantially exceeds human cognitive performance. Philosopher Nick Bostrom, whose work helped popularize the term, defines a superintelligence as an intellect that is “much smarter than the best human brains in practically every field.” Nick Bostrom The word therefore did not traditionally mean artificial intelligence in general.

It referred to a possible future state beyond ordinary AI and, depending on the framework, beyond human-level general intelligence. AI companies have also used the term in this narrower way. OpenAI’s 2023 “Superalignment” initiative, for example, explicitly discussed the challenge of aligning future superintelligent AI systems with human intent. OpenAI So there are now two meanings. In U.S. executive-branch terminology, SI is effectively the new name for AI. In much of the technical literature, superintelligence still refers to a much more capable hypothetical class of system.

What the executive order actually changes

The order is significant symbolically, but it does not instantly redefine the term for the scientific community or private industry. Section 3 of the order makes this especially clear. For the purposes of implementation, “Super Intelligence” and “SI” are initially defined using the existing statutory U.S. definition of artificial intelligence. The order then gives the Assistant to the President for Science and Technology 60 days to propose legislation for a new federal definition. The White House

In other words, the terminology has changed first. The legal definition may evolve later. The order also applies specifically to the executive branch. Universities, international organizations, private companies and researchers are not required to adopt the new language.

Why terminology matters

Technology names are not merely cosmetic. They influence how people understand what a technology is capable of doing. “Artificial intelligence” is already an umbrella term. It includes systems ranging from recommendation engines and computer vision to language models and autonomous agents. “Superintelligence,” historically, carried a much stronger implication: intelligence beyond the human level. Collapsing those meanings could make public debate more confusing. A narrow customer-service classifier and a hypothetical system exceeding human cognition across virtually every important domain are not the same technology, even if both sit somewhere inside the broader history of AI.

But terminology has always changed

There is also a broader historical point. The vocabulary of AI has never been perfectly stable. “Machine learning,” “deep learning,” “foundation models,” “generative AI,” “AGI” and “agents” all overlap in ways that can confuse non-specialists. Even “artificial intelligence,” coined in the 1950s, covers a much larger technical field today than its early researchers could realistically have anticipated. The White House’s argument is essentially that the old label undersells the present technology. Critics argue that adopting an existing technical term for a much broader category creates unnecessary confusion. Reporting from Axios and other outlets has highlighted exactly that clash between the administration’s new usage and the established research meaning. Axios

Those are two different arguments. One concerns branding. The other concerns taxonomy.

Is today’s AI already superintelligent?

Under the traditional technical definition, there is no established consensus that current AI systems qualify as superintelligence. Modern frontier models can outperform humans on some narrow or structured tasks while remaining surprisingly unreliable on others. Stanford’s 2026 AI Index calls this the jagged frontier of AI capability. Models can produce extraordinary results in mathematics or coding while still failing at apparently simple tasks. Stanford HAI

That unevenness matters. Being superhuman at selected tasks is not the same thing as possessing intelligence that greatly exceeds humans across almost every domain of interest.

Mindzy perspective

For companies actually deploying AI, the naming debate is interesting but secondary. Whether the label is AI or SI, businesses still need to answer the same practical questions. What can the system reliably do? What information can it access? What actions can it perform? Where should humans remain in control? What infrastructure does it require? Terminology can change rapidly. Architecture changes much more slowly. For that reason, Mindzy will continue to describe technologies as precisely as possible: models, AI systems, agents, AI Employees, software and compute.

The more powerful these systems become, the more important that precision becomes.

Key takeaways

• The U.S. executive branch now uses “Super Intelligence” as an administrative replacement for “Artificial Intelligence.”
• AI research has historically used superintelligence for systems that substantially exceed human cognitive performance.
• Leaders should separate political terminology from technical capability when evaluating AI systems.

Sources

1. The White House
2. Nick Bostrom
3. OpenAI
4. Stanford HAI