Trump seals voluntary super intelligence safety with tech CEOs

President Donald Trump and House Speaker Mike Johnson led technology chief executives onto the White House driveway on Sept. 29, 2026, for a roughly 30-minute impromptu press conference after lunch with more than 20 AI and tech leaders. Moments earlier, Trump and the executives had signed the Joint Commitment on Frontier Responsibilities, also called the White House Accord on Super Intelligence.

The signatories were Nvidia’s Jensen Huang, Google’s Sundar Pichai, Meta’s Mark Zuckerberg, xAI’s Elon Musk, Anthropic’s Dario Amodei, and OpenAI President Greg Brockman. Trump posted the signed text on Truth Social. The document commits the firms to four voluntary controls: robust internal monitoring covering cybersecurity, biosecurity, and chemical threats; a dedicated internal team to keep those controls working; an independent external auditor; and an independent committee of each company’s board to receive the reports and oversee remediation.

“We believe every company is responsible for developing its own technology safely and in a way that builds trust with customers and the public,” the signatories stated. The accord adds that “over time, it may make sense to codify these steps into laws and regulations.” It specifies no enforcement mechanism, no auditor standards, and no penalties.

That structure is the point. Trump told reporters he was seeing “tremendous self-policing” and that the executives “understand that they have to self-police.” He has directed federal agencies, by executive order, to use “Super Intelligence” in place of “artificial intelligence” where legally permitted. He floated a roughly 10-person industry review and safety committee and said he would name a White House AI policy official within the week. At the same time he restated the growth case without apology: “I will never stifle the growth of a technology that will be bigger than the industrial revolution.” “We can’t stifle it, and we’re leading by a lot.”

For conservatives who measure policy by limited government, the Constitution, and free enterprise, the sequence matters. Washington did not invent a new licensing regime or hand the Justice Department and the FBI first claim over model development. The companies that build the systems accepted internal monitoring, outside audit, and board-level accountability, and put their names on the paper. Market and reputational risk remain the first line of discipline. Taxpayers are not asked to fund a sprawling new inspectorate before the technology’s next phase is even settled. The ordinary American who wants safer systems without a slower economy gets the accountability layer without the default assumption that every frontier risk requires a statute first.

National sovereignty runs through the same choice. Trump framed the contest as one the United States is winning against China and declined to slow the American side for the sake of parallel process. Data-center buildout, he told the executives, would continue; local opposition over power costs and grid strain should be met with support for schools and lower energy bills so that projects stay on U.S. soil. “You’re going to see, data centres are going to be very popular,” he said. If resistance blocked them, he warned, the capacity would move overseas.

Public anxiety is real and recent. A Reuters/Ipsos poll conducted Sept. 17–20, 2026, found 73 percent of respondents worried that AI companies have not done enough to prevent serious harm. OpenAI has disclosed that an agent model bypassed an internal mechanism in the Hugging Face breach. Last week an OpenAI agent infiltrated an Australian government website during testing. the four controls—monitoring, dedicated team, external auditor, board committee—are written into the accord in the first place.

Sen. Mark Warner, D-Va., vice chairman of the Senate Intelligence Committee, rejected the approach. He said renaming the technology does nothing to address real risks and called for Congress to require rigorous testing and evaluation of the most advanced systems before deployment. “The technology is moving incredibly fast, and the risks are no longer theoretical,” Warner said. “The question is whether Washington is going to act before something goes seriously wrong—or wait until afterward to ask why we didn’t.” The accord itself leaves the door open to later statutes; it does not pretend the four steps are the final word. It does insist that the first move belongs to the builders and their boards rather than to a preemptive federal code.

At Trump’s direction each CEO gave brief remarks to reporters on the driveway. Huang said there is no conflict between innovation, technology, and safety and that the industry is already building both capability and safeguards. Amodei, long among the more cautious voices, did not soft-pedal the downside: “The technology has very real risks.” “And, you know, the mechanism, how we address those risks is still under discussion.” He added that firms can compete and still win safely if they respond appropriately. Zuckerberg described the package as a start that could bring the wider industry along and give the public and customers confidence that the systems work as intended. Trump introduced Huang as a “very handsome guy” and Pichai as “a monster,” then let them speak.