Frontier-model executives agree to four layers of internal controls, outside audits and board oversight as the president calls the pact morally binding and directs agencies to use the term SI.
President Donald Trump met Tuesday with leading technology executives for a White House session the administration billed as a meeting and luncheon on Super Intelligence. Afterward, he said the group had signed a voluntary document he called morally binding and that companies would police themselves as they train and deploy frontier models.
The gathering was in the East Room. House Speaker Mike Johnson and Vice President JD Vance were there. So were Nvidia CEO Jensen Huang, Tesla, SpaceX and xAI CEO Elon Musk, Meta CEO Mark Zuckerberg, Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, OpenAI President Greg Brockman, Amazon founder Jeff Bezos, Microsoft CEO Satya Nadella, Advanced Micro Devices CEO Lisa Su and Palantir CEO Alex Karp. Treasury Secretary Scott Bessent and other administration officials also attended.
The one-page paper is titled the White House Accord on Super Intelligence Joint Commitment on Frontier Responsibilities. Trump later posted it to Truth Social. Signatures shown with his own were those of Pichai, Amodei, Zuckerberg, Huang, Brockman and Musk.
Johnson described the text as a joint commitment and a statement of principles that is “voluntary on behalf of the industry.” Asked whether the deal binds anyone, Trump said, “I think it’s morally binding.” He also called it “almost like a constitution, in a way.”
The accord runs a little more than 300 words. It says every company training and deploying frontier models should keep “robust internal processes and controls” so systems behave as intended and problems get found and fixed. It then lists four layers of controls and audits.
First, firms are to “implement robust internal controls to monitor the capabilities and alignment of its models during training and deployment around areas like cybersecurity, biosecurity, and chemical threats, and to ensure that its models do not hack or access technical systems in unintended ways.” Second, they are to “empower an internal team to ensure all of the controls, monitoring, and detection are operating as intended, and that any issues are remediated.” Third, they are to “partner with an independent external auditor or evaluator to carry out independent assessments of whether the controls, monitoring, and detection are operating as intended.” Fourth, they are to “designate an independent committee of the board of directors to oversee and receive reports from the teams operating the controls and the internal and external auditors and evaluators, as well as to ensure any issues identified are remediated.”
Participating companies “will meet regularly to establish standards and best practices to improve the safety of their systems.” The last paragraph states: “Over time, it may make sense to codify these steps into laws or regulations. Regardless of whether this is required of companies, we believe that implementing these controls and audits is critical to ensuring a safe future for everyone, and each of our companies are committed to doing this.”
Trump talked with reporters after the meal and again outside the West Wing. “There’s a belief that there should be tremendous self-regulation, and we automatically have regulation with the Department of Justice, the FBI, all of that,” he said. “But the self-regulation is very important.” He added, “There was a lot of commonality in this room.”
On safety, he said, “I think I’m seeing tremendous self-policing, and they understand that they have to self-police.” Later: “These models are very complex and, again, I believe they’re going to be used for the good.” He also said, “It’s great for us, and we’re leading by a lot, and we’re going to keep it that way. It’s going to be very, very safe, and there’s a self policing, and there’s also a group policing, and it’s going to be great.”
The same day, Trump said he signed an executive order directing federal agencies to use “Super Intelligence,” or SI, in official correspondence, websites, reports and other non-statutory documents instead of artificial intelligence. “We’ve changed the name officially today,” he said. “We’ve changed the name officially to SI.” The order gives the assistant to the president for science and technology 60 days to send proposed legislative language defining SI and to say whether that definition should modify, expand or supersede the legal definition of AI.
He said he was “very close” to naming an AI czar and would do so in the next three to four days. Officials and executives also discussed a 10-person committee to “watch over the enterprise.”
Amodei, who has argued publicly for a slower frontier, said rules for the risks remain unfinished. “The mechanism, how we address those risks, is still under discussion,” he said. He also said: “We all need to work together to make sure that we can win, and we can win safely. If we do this right, if we work with the president and everyone here, we can win safely.” Zuckerberg called the accord “a start.” Musk, speaking at a separate America.gov event the same afternoon, said, “We agreed to a number of things that include joint monitoring, board special committees, just generally grading each other’s homework.”
Su said she left “very encouraged” and that the room had “a lot of optimism and a sense of responsibility.” “I mean, at the end of the day, it’s our responsibility to show the power of the technology as well as ensure that it’s very safe,” she said. Karp, speaking before the lunch, said, “The main issue that you have and I have is we have to take responsibility for the dangers we’re aware of. All of us do. And by the way, American people don’t want separate rules for tech people and for themselves.”
Trump also defended large data-center buildouts. He called them a “very positive thing” and said tech leaders “want to see communities that are safe and happy, and they’re going to make communities very, very happy. Otherwise, they’ll be forced to go overseas or other locations.”
For executives and developers, the practical takeaway is narrow and immediate. The White House is not imposing new statutory limits in this document. It is asking frontier labs to stand up model-alignment controls, an internal monitoring shop, an outside evaluator and a board committee — and to compare notes with rivals on a regular cadence. The paper itself leaves open later statutes. It does not write them.
No tracking, no middleman. Follow by RSS (nothing is collected) — or add your email to our self-hosted list.
RSS feed →