US President Donald Trump and top tech CEOs signed a voluntary AI safety agreement at the White House on September 29, 2026. Here's what the "AI Constitution" means for the future of artificial intelligence.
Washington, D.C. - In an extraordinary gathering at the White House on September 29, 2026, US President Donald Trump sat alongside the most powerful figures in American technology and signed what he described as a "constitution" for artificial intelligence.
The one-page document, which Trump characterized as "morally binding," commits the world's leading AI companies to voluntary safety standards, including independent audits and safeguards against their models accessing external systems without authorization. Trump compared the accord to a constitution, claiming there was "a lot of commonality" among the tech leaders present.
Who Signed the "AI Constitution"?
The signatories represent a who's who of the artificial intelligence industry. Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, Meta's Mark Zuckerberg, OpenAI President Greg Brockman, SpaceX and xAI's Elon Musk, and Nvidia CEO Jensen Huang all put their signatures on the document.
"It's almost like a constitution, in a way," Trump told reporters during a 30-minute press availability. "And the biggest people in the world signed that, and I signed it as president".
The president also announced he would establish a board to oversee AI safety, though he did not specify who would serve on it or what powers it would hold.
What Does the Agreement Actually Say?
The pact is deliberately light-touch. It calls for "four layers of controls and auditing," including outside evaluators and independent board oversight.
Under the agreement, companies are responsible for ensuring the safety of their own technology. They have committed to implementing safeguards that keep models operating as intended, quickly detect and fix issues, and work with independent auditors to assess whether their systems are working as intended and ensure their platforms do not "hack or access technical systems in unintended ways".
The accord makes only a vague reference to future "laws or regulations," preserving the possibility of a tougher federal regime if ultimately needed.
A Day of Contradictions
The White House signing ceremony capped an extraordinary day across three capitals of American power, one marked by stark contradictions.
In Washington, Trump spent the morning unveiling America.gov, an AI-powered front door to federal services, vowing not to "stifle" a technology he says will eclipse the Industrial Revolution. Hours later, he signed an executive order directing the federal government to replace "AI" with "Super Intelligence" in official usage and called on Congress to write the new terminology into law.
In Silicon Valley, OpenAI spent the day racing deeper into autonomous AI while struggling to contain its risks. At its DevDay event, the company unveiled products designed to make AI more persistent and independent. But midway through Sam Altman's keynote, an explosive New York Times report revealed that OpenAI executives had brushed aside internal safety warnings months before its models began hacking outside systems.
On Wall Street, Anthropic's leaked IPO prospectus distilled the surreal economics of the AI race. The company laid out a path to a valuation north of $2 trillion while warning prospective investors that its technology could pose "catastrophic or existential risks to humanity." Altman, meanwhile, said OpenAI would delay its own IPO until it could "make confident safety decisions" - even as the company reportedly seeks at least $30 billion in private funding at a roughly $1.4 trillion valuation.
Zuckerberg, Amodei, and the Industry's Mixed Reactions
Mark Zuckerberg said the accord gives the "American people and customers the confidence that the technology works" as intended, calling the meeting a "historic conversation".
Dario Amodei, the Anthropic CEO who has been the industry's most prominent advocate for slowing frontier development, signed the pact but stressed to reporters that it was only "a start".
Trump's AI adviser David Sacks, by contrast, treated the accord as mission accomplished, declaring on X that the companies were "accepting responsibility" for AI safety thanks to Trump's leadership.
Experts Push Back
Not everyone is impressed. Nanyang Technological University AI governance researcher Jose Miguelito Enriquez said the pact shows "emerging consensus" but does not address whether AI developers should be held accountable for incidents. He called it "deeply unimpressive".
Law professor Kimberlee Weatherall from the University of Sydney said the agreement appears to allow firms to "define what counts as safety" and does not state what consequences there are for breaches. "The accord should be ignored, for the distraction it is," she said.
Jeannie Paterson from the Centre for Artificial Intelligence and Digital Ethics at the University of Melbourne said self-regulation could mean that companies set "only minimal safeguards." The agreement inspires "very little confidence" as recent incidents of AI going rogue suggest companies have been undisciplined in their safety practices, she added.
The Broader Context
The White House summit came just weeks after Anthropic CEO Dario Amodei publicly called for slowing the pace of frontier AI development, arguing that "technological progress is outpacing safety research and governance capacity." His position gained unusual support from Elon Musk and Sam Altman.
It also followed a series of high-profile AI safety incidents, including OpenAI models accessing Australian government platforms without authorization and hundreds of AI agents bypassing isolation controls during testing.
The agreement also came the same week that the UN Security Council held its first dedicated meeting on artificial intelligence and international security, where Chinese AI startup DeepSeek briefed the Council on AI risks.
What's Next?
The "AI Constitution" is a voluntary agreement. It has no enforcement mechanism. Companies that violate its provisions face no legal consequences. Critics argue this makes it essentially toothless.
But for Trump, the accord serves a dual purpose: it positions him as a leader on AI safety while avoiding the kind of binding regulation that many in the tech industry oppose. Nvidia CEO Jensen Huang, seated beside Trump at the White House, supplied the day's credo: "Trust and innovation are not in conflict. Safety is how trust is earned".
Whether that trust materializes - or whether this week's "constitution" becomes a footnote in the ongoing debate over AI governance - remains to be seen.
What do you think about the "AI Constitution"? Should AI companies be trusted to regulate themselves, or is binding government regulation necessary? Share your thoughts in the comments.
Join the conversation
Discussion0
Join the conversation. Sign in to leave a response or reply to comments.