Trump and AI leaders sign voluntary frontier-AI pledge that misspells 'United States'

Trump and AI executives sign a frontier-AI commitment

President Donald Trump and senior figures from the AI industry signed a voluntary agreement on Tuesday committing to stronger controls and safety measures as artificial intelligence development races ahead. According to TechCrunch, the document — a "Joint Commitment on Frontier Responsibilities" — was announced following a White House meeting attended by Meta's Mark Zuckerberg, Nvidia's Jensen Huang and Anthropic's Dario Amodei, among other AI leaders.

The announcement did not go off cleanly. In a photo of the signed agreement that Trump posted to Truth Social, "United States" is misspelled as "Unites States," an error that sits directly beneath the president's signature. TechCrunch points out that the other signatories and their companies are spelled correctly. Matt Novak, a reporter at Gizmodo, was among the first to flag the mistake, and TechCrunch says it has contacted the White House for comment.

What the pledge covers

As TechCrunch describes it, the pledge sets out how companies that train and deploy frontier models should "ensure that their technology behaves as intended and that any issues are promptly identified and resolved." The measures include an independent board to provide oversight, along with internal controls at each company.

Crucially, the commitment is voluntary. TechCrunch reports that there appear to be no legal or regulatory consequences if a signatory breaks its terms, meaning the pledge functions more as a statement of shared intent than a binding rule.

Speaking at a press conference after the meeting, Zuckerberg characterized the accord as a "start" that the "whole industry could come to," and described the discussion as "sort of a historic conversation."

A compromise after a public split over speed

The signing follows a notable rift within the AI world. Weeks earlier, according to TechCrunch, OpenAI's Sam Altman, Anthropic's Amodei and Elon Musk had each called for slowing the pace of AI development — calls that both Trump and Nvidia's Huang rejected. Trump has previously framed the race in blunt terms: "We're the most sophisticated country in the world, and frankly I want to keep it that way because whoever wins AI, wins."

Against that backdrop, the pledge reads as a middle-ground outcome: language about safety and oversight, but no enforceable obligations that would slow anyone down.

Government told to say "superintelligence"

In a separate move reported by TechCrunch, Trump is using an executive order to require all government departments and agencies to replace the phrase "artificial intelligence" with "superintelligence," or "SI," in their official phrasing. The report does not detail how the renaming will be carried out across agencies.

Why it matters

The pledge is the clearest signal yet of how the administration prefers to govern frontier AI: through voluntary, industry-backed commitments rather than regulation with teeth. That approach preserves development speed — in line with Trump's stated view that whoever wins AI wins — but it also means the safeguards rest entirely on goodwill, since no signatory faces penalties for walking away. The typo itself is trivial, yet as TechCrunch observes, it offers an awkward hint at how much care went into a document meant to anchor national AI policy. And the mandated shift toward "superintelligence" in government language suggests that how the technology is framed is now as much a policy decision as the rules around it.