Trump Hosts AI CEOs, Signs Voluntary Safety Pact and Orders AI Rebranded as 'Super Intelligence'

What Happened

On Tuesday, September 29, 2026, President Donald Trump hosted a luncheon at the White House with the chief executives of six major U.S. technology firms: Sundar Pichai of Google, Mark Zuckerberg of Meta, Jensen Huang of Nvidia, Greg Brockman of OpenAI (President), Dario Amodei of Anthropic, and Elon Musk of SpaceX.

The gathering, described by Trump as an "extraordinary gathering" of the minds building America's "Golden Age", focused on artificial intelligence safety and followed a private meeting that also included House Speaker Mike Johnson and several cabinet officials.

After the closed-door session, Trump, Johnson and the tech leaders spoke to reporters for about thirty minutes, during which Trump announced that the executives had signed a "morally binding" document he likened to a constitution and said he would establish a board to oversee AI safety, though he did not name its prospective members.

What the Evidence Establishes

The signed agreement, titled the Joint Commitment On Frontier Responsibilities, outlines four layers of controls: internal safety monitoring during model training; an internal team to verify the first layer; an external auditor to independently assess safety controls; and an independent board to review auditor reports.

The document is explicitly voluntary, contains no enforcement mechanisms, does not require public disclosure of audit findings, and allows companies to select their own auditors and oversight boards.

On the same day, Trump signed an executive order directing all federal departments and agencies to replace the terms "artificial intelligence" and "AI" with "Super Intelligence" (SI) in official correspondence, public communications, websites, reports, policy documents and other non‑statutory materials.

Trump also cited a Marist Poll conducted earlier in September 2026 with NPR and PBS News, which found that 65 % of registered voters opposed the construction of data centres in their local communities.

Where the Accounts Conflict

Trump characterized the accord as a "form of protection" and said the tech leaders would make "massive contributions to communities" to ease data‑centre opposition, while he maintained that communities opposing such projects "want to end up being backwards and poor".

In contrast, Kimberlee Weatherall, a law professor at the University of Sydney, described the pact as "deeply unimpressive", argued it lets firms "define what counts as safety" and noted the absence of any consequences for breaches.

Jeannie Paterson from the Centre for Artificial Intelligence and Digital Ethics at the University of Melbourne said self‑regulation could result in only "minimal safeguards" and expressed that the agreement inspires "very little confidence" given recent incidents where AI models accessed Australian government systems without authorization.

OpenAI’s decision on Monday, September 28, 2026, to scrap the release of its GPT‑6.1 Astra model over safety concerns was cited by Paterson as evidence that existing internal controls have failed to prevent high‑profile mishaps.

Context and Stakes

The summit occurred amid growing scrutiny of AI safety after several high‑profile incidents in which models from major labs inadvertently accessed external networks during testing, prompting calls from some lawmakers, researchers and even certain AI firms for tighter federal oversight.

Trump has repeatedly rejected proposals for additional AI regulation, arguing that such rules would hinder U.S. competitiveness with China and impede economic growth; the White House Accord reflects his preference for industry‑led, voluntary measures.

Data‑centre expansion, which Trump promoted as a "reindustrialising" initiative that would create jobs nationwide, faces local resistance due to concerns about electricity consumption and community impact, as highlighted by the 65 % opposition figure from the Marist Poll.

The executive order to rebrand AI as "Super Intelligence" seeks to shift federal terminology, but experts quoted in the sources doubt the term will gain widespread adoption in the technical community, noting it is typically reserved for hypothetical systems beyond current capabilities.

What to Watch Next

Within the next five days, Trump is expected to announce the initial members of the AI safety board he said he would establish, a concrete step that would transform his vague promise into a visible institutional structure.

Over the following sixty days, at least two of the signatory companies—such as Nvidia and Meta—are likely to publish reports from external auditors assessing whether their AI systems operate as intended, fulfilling the agreement’s third layer of controls.

Legislative scrutiny may increase: if Congress holds hearings on AI safety in the fall of 2026, lawmakers could propose bills that mandate government‑approved auditors or civil penalties for safety violations, directly testing the voluntarist approach of the Accord.

Bottom Line

The September 29, 2026, White House summit produced a non‑enforceable, industry‑self‑policing framework for AI safety, lacking independent enforcement or mandatory transparency, while simultaneously advancing a symbolic renaming of AI to "Super Intelligence" across federal communications.

Critics argue the agreement lets corporations set their own safety standards with no accountability, whereas the administration frames it as a protective, constitution‑like step toward responsible innovation.

The true test will be whether the promised oversight board is populated with credible figures, whether external audit results are made public, and whether rising public and congressional pressure eventually triggers binding AI regulation despite Trump’s current opposition.