News

Trump’s New Name for AI Collides With What 'Superintelligence' Actually Means

• By John K. Waters
• 09/28/2026

At first, it sounded like a branding exercise. President Donald Trump stood before the United Nations General Assembly and announced that artificial intelligence would henceforth be called “super intelligence.” He disliked artificial, he said, because it made the technology sound fake.

Online, the jokes arrived quickly. Facebook cofounder Dustin Moskovitz posted on Bluesky, “ASI is officially here, but it’s all much dumber than we thought it would be.” Cybersecurity commentator Pwnallthethings imagined an exhausted OpenAI lawyer changing the company’s name to “OpenSI” in IPO paperwork.

Then the joke acquired a diplomatic itinerary. Three days after Trump’s speech, the White House said the United States and China had agreed to call their new bilateral AI talks the "Super Intelligence Dialogue." It also said the countries would establish a channel for reporting incidents involving the technology. A presidential preference had moved from the podium into the name of a diplomatic forum.

The distinction is bigger than it sounds. In AI research, superintelligence generally describes a prospective system far more capable than humans across a broad range of cognitive tasks. Trump is using “super intelligence” as a replacement name for AI as a whole. If that usage spreads through government, the same words could refer both to the tools people use today and to systems that researchers say would pose fundamentally different challenges.

A Name with a Job Already
In his September 22 address, Trump rejected what he called a “globalist scheme of control” for artificial intelligence, adding that it would be “hereinafter officially called ‘Super Intelligence.’” The speech linked the new name to his argument that the United States should lead the technology’s development without international oversight.

The change quickly acquired an administrative life. The Associated Press reported that a State Department official instructed employees in its international organizations bureau to replace “Artificial Intelligence” with “Super Intelligence” in documents and positions, and to use “SI” in remarks and press materials. AP said it was unclear whether other State Department bureaus or federal agencies had received similar instructions.

On social media, some people treated the announcement as another exercise in political renaming. Social Media Lab compared it to the “freedom fries” episode. Moskovitz’s joke went further: it worked because anyone familiar with AI knows that calling today’s systems superintelligent is a category error. They can be impressive and useful while remaining unreliable in ways that matter.

Superintelligence is not an empty marketing term waiting for a new definition. In a 2023 essay on its governance, OpenAI described it as future AI “dramatically more capable than even AGI,” or artificial general intelligence. The company argued that such systems might eventually need international coordination and oversight beyond what it proposed for less capable models. A research paper on superintelligence strategy similarly defines it as AI vastly better than humans at nearly all cognitive tasks. These describe a possible future capability, not a claim that every current AI system has reached it.

Calling a chatbot, a fraud detection tool, and a hypothetical system that surpasses human experts by the same name makes it harder to say which capabilities a warning or policy is actually about. The effect on policy is an inference, not an established consequence of the name change. But precision matters most when governments decide what to measure and when to intervene.

The U.S. government already has a broad legal definition of artificial intelligence. Federal law describes it as machine-based systems that make predictions, recommendations, or decisions for human-defined objectives. And Trump’s own June executive order on advanced AI does not treat every AI model alike: It calls for benchmarking advanced cyber capabilities and setting a threshold for designating a “covered frontier model.” The technical and legal machinery depends on distinctions that a blanket rename does not supply.

Two Accounts of One Conversation
The new terminology became more consequential after Trump met Chinese President Xi Jinping. In its September 25 fact sheet, the White House said both countries had agreed to use “super intelligence” instead of “artificial intelligence.” It announced a U.S.-China Super Intelligence Dialogue to discuss benefits and risks, with another exchange planned by November, as well as a bilateral channel for incidents. Those are proposed mechanisms for communication between the two leading AI powers, whatever name appears on the door.

Trump reinforced the claim in a Truth Social post, writing that Xi “seemed to like” the name and that others at the state dinner agreed. That is Trump’s account of their reaction, not an independently confirmed endorsement from Xi or the executives in attendance.

China’s published account of the same meeting tells a more limited story. It says Xi called for continued dialogue on AI, discussion of its risks and benefits, and efforts to prevent misuse. It says AI must remain under human control. The account does not mention an agreement to adopt “super intelligence” or name a new dialogue that way. That omission does not prove China rejected the wording. It does mean the White House’s claim of shared terminology is not reflected in China’s public account.

There is an irony in what the two governments publicly describe. Trump has applied a term associated with extraordinary future capability to the entire field. China’s account, meanwhile, emphasizes keeping AI under human control. Both sides describe a reason to keep talking. Their published accounts leave open whether they mean the same category of technology when they say, “super intelligence.”

A name alone cannot change what a model can do. Nor does the State Department instruction rewrite the definition in federal law. But official language shapes the questions governments ask, the agreements they announce, and the risks they explain to the public. Trump’s rebrand is now attached to a real diplomatic effort. The test will be whether that effort develops clear ways to identify dangerous capabilities, communicate incidents, and distinguish present systems from the far more powerful ones the name already evokes.