The United States is pushing governments around the world to replace the term “artificial intelligence” with “super intelligence,” turning a dispute over two words into an international policy issue.
The Trump administration argues that the new name better represents the capabilities of modern AI systems and could help position the United States as a leader in a technology it considers essential to economic growth.
But scientists see a fundamental problem: superintelligence already has an established meaning in AI research, referring to hypothetical systems that greatly exceed human intelligence across nearly every cognitive domain. No publicly demonstrated AI system has met that standard.
The dispute reached international financial negotiations on October 15, when U.S. Treasury Undersecretary for International Affairs Erin Browne confirmed that Washington was pressing other countries to adopt the terminology in an International Monetary Fund statement.
According to Reuters‘ report on the Treasury Department’s proposal, the United States wanted the change to extend beyond federal government documents and become part of international discussions about technology and economic development.
The campaign has three overlapping motivations: changing how the public perceives AI, reinforcing America’s leadership in technology, and influencing the language used by international institutions.
For researchers, however, the proposed terminology risks confusing today’s increasingly capable AI tools with a much more advanced form of intelligence that remains unproven.
Why Trump Wants to Replace Artificial Intelligence With Super Intelligence
President Donald Trump’s objections begin with the word “artificial.” During his September 22 address to the United Nations General Assembly, Trump argued that the term makes AI sound less real or important than it is.
He announced that U.S. government documents would instead use “super intelligence,” presenting the change as a more accurate description of a technology he believes could transform the economy.
As Axios reported on Trump’s UN speech, the president also criticized international efforts to regulate AI and framed American technological leadership as a strategic priority.
On September 29, Trump formalized the change through an executive order directing federal agencies to replace references to artificial intelligence with the terms “Super Intelligence” and “SI.”
The order describes existing frontier AI systems as more capable than technologies that merely imitate individual aspects of human intelligence.
The administration’s definition therefore differs from the meaning commonly used by researchers.
Rather than treating superintelligence as a future level of capability, the White House is using it as a new name for the broad category of technology already known as AI.
That difference is at the center of the controversy.
The administration’s argument is largely about how AI should be presented. It wants a term that emphasizes capability, technological progress and economic opportunity rather than something that sounds artificial or manufactured.
The scientific objection is that changing a technology’s name does not change what it can actually do.
The U.S. Is Pressuring International Institutions to Adopt the New Name
The renaming campaign is no longer limited to speeches and federal government communications.
During the IMF and World Bank annual meetings in Bangkok, U.S. Treasury officials pushed for the term “super intelligence” to appear in a statement being negotiated by the International Monetary and Financial Committee.
The committee plays an important role in setting the direction of international monetary and financial cooperation.
Treasury Undersecretary Erin Browne told Reuters that the administration considered the terminology a priority and wanted other countries to follow Washington’s example.
The United States also holds the rotating G20 presidency, giving it another platform to promote its preferred terminology.
The effort reportedly surprised some other IMF members.
According to Reuters, officials discussed the possibility of avoiding both terms and referring more generally to “new technology” if negotiators could not reach an agreement.
Browne declined to say whether Washington would refuse to support the statement unless the wording changed.
The disagreement illustrates how terminology can become part of international economic negotiations.
For the United States, promoting “super intelligence” could reinforce its preferred message about innovation and technological leadership.
For other countries, accepting the term could create confusion in policy documents, particularly when existing laws, research standards and regulatory frameworks already use artificial intelligence.
The distinction is more than cosmetic. Governments rely on consistent definitions when developing regulations, coordinating research and evaluating emerging technologies.
What Does Superintelligence Actually Mean?
In scientific discussions, superintelligence has traditionally referred to something far more advanced than today’s AI systems.
The concept was popularized by philosopher Nick Bostrom in his 2014 book Superintelligence: Paths, Dangers, Strategies.
Bostrom described a hypothetical form of intelligence capable of outperforming humans across virtually all important cognitive activities.
That could include scientific discovery, strategic planning, complex problem-solving and other intellectual tasks.
The concept is distinct from artificial general intelligence, or AGI, which is commonly used to describe systems capable of performing a wide range of cognitive tasks at roughly human-level competence.
Superintelligence goes further by suggesting capabilities substantially beyond those of humans.
As Nature explained in an October 7 scientific analysis, researchers generally regard superintelligence as an aspirational category rather than an achievement demonstrated by existing technology.
Current AI models can outperform humans on selected tasks, including some programming, mathematics and scientific evaluations.
But strong performance in particular areas does not establish that a system possesses broad, consistently superior intelligence.
Modern AI systems can still make factual mistakes, misunderstand instructions, produce unreliable conclusions and struggle with tasks requiring sustained real-world judgment.
There is also no universally accepted scientific test that establishes when a system becomes superintelligent. That uncertainty makes the administration’s terminology especially controversial.
ALSO READ: Human-Level AI by 2040?
Why AI Researchers Object to the New Terminology
The scientific criticism is not simply that the White House has chosen an unfamiliar name.
It is that the proposed replacement could erase an important distinction between current technology and a hypothetical future capability.
Melanie Mitchell, an AI researcher at the Santa Fe Institute, told Nature that concepts such as human-level intelligence and superintelligence lack precise, universally accepted definitions.
Her concern reflects a wider difficulty in AI research: scientists still disagree about how to measure general intelligence and which capabilities should count.
The absence of a single definition does not mean every use of the term is equally informative.
Calling all AI systems superintelligent could imply that today’s technology has already achieved capabilities that researchers are still trying to understand and develop.
That could create several problems.
First, it may encourage exaggerated expectations about what AI can reliably accomplish.
Second, it could make public discussions about AI safety more confusing by applying the same term to current systems and hypothetical systems with much greater capabilities.
Third, it may weaken distinctions used by researchers and policymakers to assess different levels of technological risk.
An AI tool that summarizes documents and a hypothetical system that surpasses humans across nearly every intellectual domain do not necessarily present the same capabilities or risks.
Using the same label for both could make those differences harder to communicate.
Nature’s analysis also noted that researchers do not agree on every aspect of the superintelligence concept.
Nevertheless, the term’s longstanding association with beyond-human capabilities makes its use as a general replacement for AI scientifically contentious.
The White House May Also Be Trying to Change Public Perception
The renaming effort comes as Americans express growing concerns about artificial intelligence.
AI companies have promoted the technology as a way to improve productivity, accelerate scientific discovery and create new economic opportunities.
At the same time, consumers and workers have raised concerns about job displacement, privacy, misinformation and the reliability of AI systems.
Darrell West, a technology policy researcher at the Brookings Institution, told Nature that the administration’s campaign could be an attempt to improve public acceptance of AI by giving it a more positive name.
West was skeptical that a change in terminology would be enough to shift attitudes.
The argument highlights a possible political motivation behind the rebranding.
“Artificial intelligence” has become associated with both technological progress and fears about automation.
“Super intelligence,” by contrast, places greater emphasis on extraordinary capability.
That may fit the administration’s message that American innovation should be encouraged rather than constrained by excessive regulation.
But the new term could also have the opposite effect.
For people already concerned about advanced AI, describing current systems as superintelligent may increase fears that the technology is becoming more capable than humans can control.
A more ambitious name does not resolve questions about job security, misinformation, data protection or AI accountability.
Those concerns depend on how the systems operate and how they are governed, not on what governments call them.
Elon Musk Embraces the Change, but OpenAI Is Keeping Its Name
The campaign has received a mixed response from technology executives.
Elon Musk publicly supported the terminology change and subsequently moved to rename his company’s AI division from SpaceXAI to SpaceXSI.
Microsoft CEO Satya Nadella also adopted the new terminology in a public statement about the risks posed by advanced AI models.
OpenAI CEO Sam Altman took a different position.
According to Business Insider’s reporting on the industry’s response, Altman said OpenAI had no plans to change its name, although the company would continue discussing superintelligence as a technological goal.
The contrasting responses reveal an important distinction.
AI companies may agree that superintelligence is a worthwhile research objective without agreeing that today’s AI systems should already be called superintelligent.
For companies such as OpenAI and Anthropic, the difference matters because developing systems with capabilities beyond human intelligence remains part of their long-term research ambitions.
Calling existing AI technology superintelligence could blur the distinction between what has already been achieved and what researchers hope to develop.
ALSO READ: Meta Says Its New AI Models Are on Par With OpenAI and Anthropic
The IMF Avoided Both AI and Super Intelligence in Its Final Statement
The international dispute produced a revealing outcome on October 17.
Despite Washington’s push for the new terminology, the IMF’s International Monetary and Financial Committee did not adopt “super intelligence” in its final statement.
Instead, the committee used “advanced intelligence technologies,” avoiding both the administration’s preferred term and the conventional phrase “artificial intelligence.”
According to Reuters‘ follow-up reporting on the IMF negotiations, committee chair Mohammed Al-Jadaan said members had agreed on the language.
The wording appears in the official October 17 IMFC chair’s statement, which discusses both the economic opportunities and financial risks associated with advanced intelligence technologies.
The outcome is significant because it shows that the United States has not yet secured international acceptance of its preferred terminology.
It also illustrates the difficulty of changing established technical language through diplomatic pressure.
International institutions must consider how terminology affects policy discussions, economic analysis and communication among member countries.
The IMF’s compromise avoided taking a position on whether existing AI systems should be described as superintelligent.
However, the outcome does not necessarily end the administration’s campaign.
The United States can continue promoting the terminology through other diplomatic channels, international forums and federal policy documents.
A New Name Cannot Settle the Debate Over AI’s Capabilities
The Trump administration’s effort to rename AI has turned a scientific distinction into a question of political messaging and international influence.
For Washington, “super intelligence” communicates technological ambition, economic opportunity and confidence in American innovation.
For researchers, the term carries a different meaning: a hypothetical level of intelligence that existing systems have not been shown to achieve.
Both positions reflect different objectives, but they are not interchangeable scientific claims.
The administration can direct federal agencies to use a particular name. It can also encourage companies and international organizations to follow.
What it cannot establish through terminology alone is that today’s AI systems have reached superhuman general intelligence.
The IMF’s decision to use neutral language suggests that other governments are not necessarily ready to accept Washington’s preferred definition.
The larger question is whether changing the name of artificial intelligence will change how people understand, trust or regulate it.
For now, the answer remains uncertain. AI is becoming more capable, and governments are competing to shape its development. But calling it superintelligence does not make it so.
