US President Donald Trump and the heads of leading technology companies signed a voluntary accord on artificial intelligence safety at the White House on Tuesday. The agreement relies on companies policing themselves and does not create new legal rules. Trump called the document "morally binding" and compared it to "a constitution, in a way."
The accord was signed by Trump, Google chief Sundar Pichai, Anthropic chief Dario Amodei, Meta's Mark Zuckerberg, OpenAI President Greg Brockman, Nvidia chief Jensen Huang and SpaceX's Elon Musk. It commits firms to "robust internal controls", to working with an independent external auditor to check whether those controls work, and to setting up a board-level committee in each company to review auditors' reports. It says it may "make sense to codify these steps into laws and regulations" over time. Trump said he would name a roughly 10-person committee to oversee the effort, and someone to supervise the agreement in coming days, without saying who. Some of the measures reflect steps the companies already take or have promised before.
The summit took place as concern grows about AI's risks. Some industry figures have warned the technology could one day outsmart humans. Amodei, among the most outspoken advocates of caution, said the technology has "very real risks" and that how to address them is "still under discussion." Zuckerberg called the meeting a "historic conversation" and said the accord would give the public confidence the technology works as intended.
Critics were less convinced. Kimberlee Weatherall, a law professor at the University of Sydney, called the pact "deeply unimpressive" and said it lets firms "define what counts as safety" without spelling out consequences for breaches. Jose Miguelito Enriquez of Nanyang Technological University in Singapore said it shows an "emerging consensus" but leaves open whether developers should be held accountable for incidents. Gary Marcus, an AI researcher and professor emeritus at New York University, told PBS NewsHour that self-regulation is "certainly not enough" and called for independent evaluation of AI products, much as drugs are assessed. He argued the government already delays some systems under a framework that is not public.
Trump had said beforehand that he would "never stifle the growth" of a technology he expects to be bigger than the Industrial Revolution. He framed AI as a two-way race with China, saying there is "nobody in third place." The two countries agreed last week to set up a channel for handling AI-related incidents during Chinese President Xi Jinping's visit to Washington, but Trump said he did not want to cooperate much further.
The president also defended the rapid build-out of data centres, the large computing facilities that power AI. They have drawn bipartisan local opposition over electricity use and community impact. A Marist Poll found 65% of registered voters opposed them being built nearby, an issue ahead of November's midterm elections. Trump said companies would provide more support to local schools or act to reduce energy costs. Separately, he signed an executive order telling federal agencies to use the term "Super Intelligence" instead of artificial intelligence.
The accord matters because it shows the White House's preferred approach to AI: industry-led oversight rather than binding rules, even as debate over enforceable safeguards continues.