Trump signs 'super intelligence' order as tech executives join White House accord

By Splitfeed EditorialLast updated Sep 30, 2026, 11:40 UTC

On September 29, 2026, President Trump signed an executive order rebranding artificial intelligence as "super intelligence" at a White House summit attended by technology executives. The same day, AI companies signed a voluntary accord with the White House on frontier model safety, committing to self-policing practices rather than accepting binding federal rules. Separately, Senators Elizabeth Warren and Richard Blumenthal sent requests for details on White House AI model evaluations, signaling early congressional interest in how the administration intends to oversee these commitments.

The rebranding and the voluntary accord arrive at a moment when the debate over how to govern advanced AI has no settled answer in Washington. The executive order itself changes little in practical regulatory terms, but the choice of language signals an administration posture that treats frontier AI as a national prestige project rather than a risk to be managed through hard law. The voluntary nature of the industry accord is the sharpest point of contention: companies are agreeing to terms they negotiated with an administration sympathetic to their growth, without an enforcement mechanism, while two prominent Senate critics are already asking how evaluations will actually work.

The Anti-AI camp has reacted with visible skepticism. Michael Zimmer, posting on Bluesky, responded to the accord with "#smh," a shorthand signal of disbelief or disappointment, capturing a broader concern that a self-policing framework gives companies credit for safety without accountability. No published reactions from the Pro-AI camp have appeared yet, though that camp would typically argue that voluntary accords allow faster, more flexible progress than legislation and demonstrate industry responsibility without stifling innovation. The Middle Ground camp has also not weighed in publicly, but observers in that space would generally be expected to welcome some form of structured safety commitment while flagging the absence of enforcement as a meaningful gap.

The argument is likely to sharpen around whatever details Warren and Blumenthal receive in response to their requests, and whether the administration discloses how frontier model evaluations are conducted and by whom. If those details reveal thin or industry-controlled processes, pressure for binding rules would probably grow; if they show genuine independent review, the voluntary framework gains more credibility.