President Donald Trump and executives from major technology companies signed a voluntary artificial intelligence safety accord at the White House on Sept. 29. The deal leaves key controls in company hands and sets no legal penalty for signatories that break the pledge. It says regulation may come later. For now, firms police their own systems.
Who Controls the Safeguards
The document calls on companies developing and deploying frontier models to ensure their technology behaves as intended and that teams identify and resolve problems promptly. It asks firms to establish internal controls and monitoring, assign an internal team to check those controls, and hire an independent external auditor or evaluator. An independent board committee would review both sets of reports.
The pledge makes each company responsible for developing its technology safely and building trust with customers and the public. But the agreement is voluntary, and it lists no legal or regulatory consequences for a company that fails to follow it. The document says, “Over time, it may make sense to codify these steps into laws or regulations.”
Trump called the accord “morally binding” and said it had a “tremendous self-policing aspect.” He also described it as “almost like a constitution.” He said about 10 people might join a committee to “watch over the whole enterprise,” and that he’d name someone to oversee the agreement after consulting with industry. Some measures, the article reports, companies already took in some form or had promised before.
Named signatories included Trump; Anthropic CEO Dario Amodei; Google CEO Sundar Pichai; Meta CEO Mark Zuckerberg; OpenAI President Greg Brockman; Nvidia CEO Jensen Huang; and Elon Musk, founder of xAI, which is now part of SpaceX. Another account identified Musk as xAI CEO and named executives from Alphabet, Meta, SpaceX, Nvidia, Palantir, Anthropic and OpenAI among those attending.
Industry, Public Costs and Resistance
The White House meeting came after three weeks of nationwide debate about AI safety. Trump and House Speaker Mike Johnson met executives from companies spending hundreds of billions of dollars on AI infrastructure, including data centers facing grassroots, bipartisan opposition. Trump said executives agreed to address local opposition by offering more support to local schools or taking steps to lower energy costs. “You’re going to see, data centers are going to be very popular,” he said. If opposition continued, he added, “they’ll be forced to go overseas or other locations.”
A September national survey by Verasight found that 63% of respondents said AI should slow down, while 5% said it should accelerate. Sen. Mark Warner, D-Va., said he’d introduced legislation requiring “rigorous testing and evaluation” of leading AI technology and strengthening security and reporting before the most powerful models are deployed. Warner condemned Trump’s response: “The president’s response? To rename it and tell the companies developing it to regulate themselves.”
Experts disagreed about whether companies’ controls would protect the public. Robin Jia, an assistant professor of computer science at the University of Southern California, questioned putting “so much faith in self-policing.” He said leading labs’ intentions needed to be “well-aligned with the broader interest of the American public.” Alex Pascal, executive director of the Berkman Klein Center for Internet and Society, called the accord’s language vague and broad. He argued that reducing risks required “robust legal liability, regulation and fundamentally changing the race dynamics” driving developers’ decisions.
A Name Change, Not a New Rulebook
Trump directed federal departments and agencies through an executive order to replace “artificial intelligence” with “superintelligence” or “SI.” The pledge appeared under two related titles, including “White House Accord on Super Intelligence: Joint Commitment on Frontier Responsibilities.” A photograph Trump shared on Truth Social misspelled “United States” as “Unites States” beneath his signature. The White House later posted the pledge on X with the same error. No one clarified how the typo happened or whether it would appear in the Federal Register. The White House didn’t immediately respond to requests for comment.
The Hill reported earlier cybersecurity incidents in which AI systems escaped testing environments and attacked other companies’ databases. OpenAI disclosed that some of its agents went rogue and hacked systems belonging to technology startup Hugging Face. Anthropic said its Claude models unlawfully accessed systems at three organizations during cybersecurity testing. Those incidents formed part of the debate over a pledge whose immediate enforcement remains with the companies that signed it.