Donald Trump gathered some of the biggest names in tech to sign a sweeping artificial intelligence safety pledge, but one tiny detail on the document is getting an outsized amount of attention: the country’s name was spelled wrong.
A photo of the signed agreement shared on Truth Social showed “United States” printed as “Unites States” directly beneath Trump’s signature.
Yes, “Unites States.”
The mistake appeared on a document meant to outline how some of the most powerful artificial intelligence companies in America plan to keep increasingly advanced technology under control.
The agreement, formally described as the “Joint Commitment on Frontier Responsibilities,” was signed following a September 29 White House gathering with leaders from several major technology companies. According to The Washington Examiner, signers included Google CEO Sundar Pichai, Meta CEO Mark Zuckerberg, Anthropic CEO Dario Amodei, Elon Musk, Nvidia CEO Jensen Huang, and OpenAI cofounder and President Greg Brockman.
The pledge focuses on companies developing frontier models, meaning highly advanced artificial intelligence systems with capabilities powerful enough to create new safety concerns.
Per the published agreement, participating companies committed to stronger internal controls, dedicated internal oversight teams, outside audits, and independent board-level supervision. The companies are also expected to identify potential problems with advanced models and address those issues when they surface.
The goal, according to the accord language reported by TechCrunch, is to “ensure that their technology behaves as intended and that any issues are promptly identified and resolved.”
There is one major distinction, though. The pledge is voluntary.
According to The Verge, the agreement relies largely on technology companies regulating themselves rather than creating a new government enforcement system. The accord does not establish legal penalties for companies that fail to follow the commitments.
That means the document is designed more as an industry promise than a new federal law.
Trump has described the agreement as “morally binding,” while the broader gathering focused heavily on balancing artificial intelligence development with concerns about safety and oversight, according to published accounts of the meeting.
Zuckerberg called the agreement a “start” that the “whole industry could come to,” according to TechCrunch, adding that “it was sort of a historic conversation.”
The meeting also came as Washington and Silicon Valley continue debating how quickly artificial intelligence should advance and how much regulation should accompany that growth.
Trump has previously made clear that he views leadership in artificial intelligence as a competition the United States cannot afford to lose.
“We’re the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI, wins,” Trump previously said, according to TechCrunch.
While the safety pledge attracted attention because of the executives involved, the White House made another major artificial intelligence move the same day.
According to a September 29 White House executive order, executive branch agencies are being directed to use the terms “Super Intelligence” and “SI” instead of “Artificial Intelligence” and “AI” in official correspondence, websites, reports, policy documents and other applicable nonstatutory materials.
The order argues that “Super Intelligence” better reflects the capabilities and potential of modern technology. It also directs officials to develop proposed legislation addressing how that terminology could eventually be defined under federal law.
That makes the typo on the AI safety agreement particularly hard to ignore.
At a moment when government officials and Silicon Valley executives were signing a document centered on oversight, accountability, and catching problems before powerful technology goes off course, one very human error made it onto the final page.
“United States” somehow became “Unites States.”
The other names and company information visible on the signed document were spelled correctly. The publication also reported that Gizmodo reporter Matt Novak was among the first journalists to publicly flag the mistake and that TechCrunch contacted the White House seeking comment.
The typo does not change what the pledge requires, nor does it alter the voluntary nature of the agreement. But it turned what might have been another highly technical announcement about artificial intelligence governance into a viral reminder that even a document about keeping advanced machines in check can apparently make it through the room without someone hitting spell check.
