Trump's "Super Intelligence" Accord Has No Legal Teeth

In Brief

President Trump brought six major AI companies together for a voluntary safety accord, then ordered federal agencies to call AI Super Intelligence. 

The pledge calls for outside checks but has no legal force; the name change governs federal language and has a different meaning from the technical term superintelligence.

What Happened

On September 29, President Trump and leaders from Google, Anthropic, Meta, OpenAI, xAI, and Nvidia signed the White House Accord on Super Intelligence. It asks companies to check their most advanced AI systems for serious risks, including dangerous cyber activity and biological or chemical misuse. Trump described the agreement as morally binding, but the companies' commitments are voluntary.

The accord calls for four layers of review: company safeguards, an internal testing team, an independent outside evaluator, and a board committee to oversee fixes. Semafor reported that Meta CEO Mark Zuckerberg shaped an early draft and Nvidia CEO Jensen Huang gathered support.

That same day, Trump issued a separate executive order directing federal agencies to use Super Intelligence and SI instead of Artificial Intelligence and AI in official messages, websites, reports, and policy documents where federal law permits. For now, the order defines that two-word phrase by reference to the existing federal definition of AI.

SmarterX founder and CEO Paul Roetzer broke down the accord and the language shift on Episode 245 of The Artificial Intelligence Show.

The Key Numbers

6 - Major AI companies named in the White House signing

4 - Layers of company safety review in the accord

September 29 - Date of the accord and separate executive order

70 - Approximate years the term artificial intelligence has existed

2023 - Year of an earlier White House voluntary AI commitment

Why the Pact and the Name Change Need Separate Scrutiny

The accord sets a safety process. Roetzer stressed that “this accord is totally separate from the name change thing.” Its commitments concern model testing only. The signers did not agree in that document to start calling their products Super Intelligence. That directive is for federal agencies.

The new name creates confusion. In AI research, superintelligence, all one word, describes a future system that surpasses humans across mental tasks. Trump's Super Intelligence, two words, is the administration's new label for what federal law currently calls AI. The order changes agency language; it does not establish that researchers have achieved the more advanced capability. "The name change means literally nothing unless you are a tech leader or government official who has to go along with this to appease Trump," Roetzer says.

A voluntary pledge depends on follow-through. The accord includes outside evaluation, but companies must carry out the checks themselves. “I don't believe the accord is a realistic or meaningful alternative to any form of regulation,” Roetzer says. He pointed to the 2023 voluntary AI commitments as a reminder that an agreement can leave accountability unresolved.

“I do actually think that bringing these leaders together is an important step, and getting them to open dialogue with each other when many of them hate each other is really important for America and for AI more broadly.”

— Paul Roetzer, founder and CEO of SmarterX on Episode 245 of The Artificial Intelligence Show

AI leaders stay mum while Trump says AI is now Super Intelligence. Roetzer watched Trump tell reporters that everyone had agreed to call it Super Intelligence while the executives stood silent. He called the scene a loyalty test: With trillions at stake, none publicly corrected the claim. Roetzer described their response as sycophancy.

SmarterX Take

The pact's strongest immediate value might be getting competitors who distrust one another in a room together. “I don't believe the accord is a realistic or meaningful alternative to any form of regulation, but it is noteworthy they did it," Roetzer says. It still leaves customers needing proof that the promised safety checks happen. Business leaders should ask who reviews each system, what they can see, and what changes after a problem is found.

Clear language also matters when teams discuss AI risk. A vendor's claim about superintelligence should prompt a simple follow-up: Does it mean the technology available today, as in the Trump executive order, or a claimed leap beyond human ability per the experts definition? Those are very different assertions, with very different implications for purchasing and oversight.

What to Watch

The first outside reviews will test the accord's value. Customers and policymakers can look for named evaluators, clear testing standards, and evidence that companies fix problems those reviews uncover. Meetings alone will reveal little about whether the safeguards work.

Federal documents will show how far the renamed AI spreads. The order covers executive-branch communications within legal limits. Its practical effect will become clearer as agencies update public materials.

How Many Companies Have the Governance to Scale AI?

Only 13% of organizations have all four governance foundations measured in the 2026 State of AI for Business Report: an AI council, a roadmap, a generative AI policy, and an AI ethics policy. That gap puts the White House debate in context. Even if technology providers improve their reviews, most customers still have substantial governance work of their own.

The report draws on responses from more than 2,100 professionals across roles, functions, and industries. It offers a benchmark for leaders deciding how to turn AI safety promises into policies, accountability, and day-to-day practice inside their organizations. Read the full report →