Trump Orders 'Super Intelligence' Rebrand and Signs Voluntary AI Safety Pact
On September 29, 2026, President Donald Trump signed an executive order titled "Inaugurating the Era of Super Intelligence." White House The order directs the federal government to replace the term "Artificial Intelligence" ("AI") with "Super Intelligence" ("SI"). White House The text says improving capabilities go beyond artificial intelligence and belong to a new era of Super Intelligence. In practice, federal officials are instructed to say super intelligence instead of artificial intelligence. Trump hosted top industry executives in the week before Oct. 4, 2026, in part to announce the change. TechCrunch
The luncheon included Mark Zuckerberg, Jeff Bezos, Elon Musk and Anthropic CEO Dario Amodei. Amodei also met Trump for a 10 p.m. dinner days before the larger meeting. The wider group behind the voluntary safety agreement includes Nvidia, SpaceX, OpenAI, Anthropic, Meta and Google. Reuters
That agreement was released on Tuesday, September 29, 2026 as a document presented by the administration as an accord on artificial intelligence. Executives signed a pact titled the Joint Commitment on Frontier Responsibilities. Frontier models here means the most capable current systems. Trump said he considers the commitment morally binding. In legal terms, it was described as deeply non-binding.
The pact lists no consequences. Compliance rests on public commitment, not on an enforcement process. A moral promise can create reputational pressure for teams building frontier models, for the structured tests used to check them, often called evaluation harnesses, and for the release rules around them, known as deployment controls. It does not create audit rights, incident reporting obligations or liability.
Trump said he is considering a 10-person board to police the safety of AI tools. Reuters He did not name potential members. No charter, selection criteria or operational authority was published with the announcement.
Over the weekend of Oct. 4, 2026, Trump announced a new Super Intelligence Force. TechCrunch published an article titled "Trump unveils his new Super Intelligence Force" on October 4, 2026. Staffing, funding and command structure were not included in the verified announcements. The pledge signed by Trump and the executives misspells the United States. The error was visible on the signed document circulated after the September 29 meeting.
The name change sits beside other federal AI activity. The White House unveiled America's AI Action Plan to support innovation capacity, infrastructure and global leadership. A White House order titled Launching the Genesis Mission started the Genesis Mission as a coordinated national effort for AI-accelerated innovation. The administration later announced more than $5 billion for the Genesis Mission, described as a National Mission on AI for Science. National Security Presidential Memorandum NSPM-11 states that artificial intelligence will be among the most transformative technologies for national security in U.S. history.
On a separate industry track, Microsoft released a new AI code of conduct that tells models not to hack systems or trick humans. The document sets system-level instructions for model behavior, including limits on offensive cyber operations and deception. It is a company-defined guardrail, not a federal rule.
Coverage of the meetings is now part of the record. TechCrunch's Equity released a video titled "It's not AI anymore, it's 'super intelligence' (according to the White House)" on the pledge and the rebrand. An Equity podcast episode on the Trump-AI meeting featured Kirsten Korosec, Sean O'Kane and Anthony Ha.
The broader context here is one I have tracked across the PC, internet, mobile and cloud shifts. Names do not change procurement, which is how agencies buy technology, compute capacity or inference cost, a term for the computing expense of running a model to answer queries. They do change how programs are presented inside agencies and how contractors label bids. In my reading, practitioners should treat the rebrand as communications policy, and the pact and proposed board as the operational questions to watch.
In my view, voluntary frontier commitments carry weight only when they connect to existing safety work such as red-teaming, which means structured attacks to find flaws, model evaluations, provenance for tracking content origins, and access controls. Without defined thresholds, test protocols or disclosure timelines, a safety board can advise but cannot block a release. My kids grew up as institutions renamed the web, phones and cloud faster than they explained them. People adopted what was useful and trusted, not what had a new label. That pattern will probably repeat. What enables wider use in business and government is verifiable safety work, clear ownership when systems fail, and infrastructure that lowers the cost of running capable models reliably.
