Trump signs AI self-policing accord and renames technology 'Super Intelligence'

By Splitfeed EditorialLast updated Sep 30, 2026, 22:40 UTC

On September 29, 2026, President Trump held a White House lunch with prominent AI executives including Mark Zuckerberg, Dario Amodei, and Jensen Huang. The meeting produced a voluntary accord described by Trump as "almost like a constitution" and "morally binding," committing participating companies to self-regulate AI development through a reported four-tier structure with independent audits. Trump also signed an executive order directing all federal documents, websites, and press releases to replace the term "artificial intelligence" with "Super Intelligence." He ruled out any joint AI governance venture with China and pledged never to "stifle" AI development. The accord's full text, published by the New York Post, went viral in part because it contained a typo reading "Unites States" instead of "United States."

The accord arrives as rogue AI incidents are reportedly rising, and as senators including Warren, Blumenthal, and Slotkin are demanding White House accounting of AI model evaluations and tech company tax breaks. Geoffrey Hinton, widely called the godfather of AI, said Trump was "ill-advised" about the technology. Critics immediately drew comparisons between this agreement and the voluntary pledges secured by the Biden administration, which were similarly non-binding. The core dispute is whether voluntary self-regulation by the companies being regulated constitutes meaningful oversight, or simply formalizes an arrangement that favors industry.

The Pro-AI camp frames the accord and the rebranding as broadly positive, with the competitive argument that working with China on AI governance would slow American companies and hand rivals an advantage, a point echoed by Fllics on the question of the US-China relationship. The Anti-AI camp is sharply critical on multiple fronts. Senator Ed Markey called the morally binding agreement meaningless and argued Congress must act rather than let Big Tech regulate itself. MeidasTouch highlighted the typo as an irony that undercut the "Super Intelligence" branding. The Alt National Park Service accounts noted that the government's own AI chat tool returned inconsistent answers on politically sensitive questions about Trump. The Middle Ground camp has not produced a clear unified reaction; accounts in that space appear more focused on testing the new government AI tool than on endorsing or condemning the accord itself.

The practical test of the agreement will come when independent audits under the four-tier framework are either conducted and published or quietly abandoned. Congressional pressure from Warren, Blumenthal, and Slotkin may force earlier disclosure about how companies are actually meeting their self-imposed commitments.