President Donald Trump says top tech firms have signed a voluntary accord to “self-police” AI development, a move aimed at calming public fears about the technology while rejecting any limits on its growth.
The meeting took place at the White House and brought together executives from some of the largest AI companies in the world. Trump told reporters: “I think I’m seeing tremendous self-policing. And they understand that they have to self-police.”
Who Signed the Accord
The accord was signed by Trump himself, along with executives from major companies:
- Anthropic CEO Dario Amodei
- Google CEO Sundar Pichai
- Meta CEO Mark Zuckerberg
- OpenAI President Greg Brockman
- Nvidia CEO Jensen Huang
- Elon Musk, founder of xAI, which is now part of SpaceX
The accord was posted on Trump’s social media site, Truth Social, late Tuesday. It focuses on four voluntary steps:
- Implementing “robust internal controls”
- Partnering with an “independent external auditor”
- Establishing a board-level committee to evaluate audits
- Opening the door to future regulation
The accord states: “Over time, it may make sense to codify these steps into laws and regulations.” Amodei said: “The technology has very real risks” and added, “The mechanism, how we address those risks is still under discussion.”
Trump suggested about 10 people would be named to a committee to “watch over the whole enterprise” and said the accord would be “morally binding.” He said he would name someone to oversee the agreement in coming days after consulting with industry.
The Critics
Not everyone agrees that self-policing is the right approach. Some actions in the accord mirror measures companies are already taking or have previously committed to.
Shri Narayanan, USC engineering professor, said the accord balances innovation with regulation and includes model safety and security improvements. Robin Jia, USC computer science professor, called putting “so much faith in self-policing” skeptical and said lab intentions must align with public interest.
Alex Pascal, Berkman Klein Center executive director, said robust legal liability, regulation, and changing developer cost dynamics are needed to reduce AI risks to Americans.
Trump’s Stance
Earlier Tuesday, Trump said the U.S. government wouldn’t limit AI development or support calls for “guardrails” or other restrictions. “I will never stifle the growth of a technology that will be bigger than the industrial revolution,” Trump said at the unveiling of an AI chatbot, America.gov.
The chatbot is intended to help Americans navigate government services. Trump administration officials celebrated a “Golden Age” fueled by what Trump calls “super intelligence.”
Secretary of State Marco Rubio highlighted the chatbot’s potential to process first-time passport applications online without an in-person visit. Trump also said executives agreed to support local schools or reduce energy costs to address grassroots, bipartisan opposition to data centers.
“You’re going to see, data centers are going to be very popular,” Trump said. If opposition continues, he suggested facilities “will be forced to go overseas or other locations.”
OpenAI CEO Sam Altman said in a TV interview that the company is increasingly focused on safety, adding: “We are pacing our progress, which includes sometimes not training a model.”
The Bottom Line
The accord is a political move as much as a technical one. Trump is leaning on corporate self-regulation to manage public fear of AI while rejecting guardrails and limits, arguing that the technology will drive economic growth.
The White House is betting that moral pressure is enough. Whether that bet pays off depends on whether the companies actually follow through and whether the public trusts them to do so.
The skepticism is real. Putting so much faith in self-policing is a gamble. The technology has very real risks, and the mechanism for addressing those risks is still under discussion.
But the accord is not meaningless. It sets a baseline of voluntary action that companies can point to when critics ask what they are doing. It also leaves the door open to future regulation, which could become more urgent if the risks materialize.
For now, the deal is on the table. The question is whether it holds.
Where the paper stands
The paper backs the narrow disclosure requirement that forces companies to reveal hidden safety failures and is against the voluntary self-regulation that lets the biggest firms hide their failures. Trump’s accord is a political move dressed up as a solution, and it does nothing to address the risk of concentration of power among the companies involved.
The accord’s voluntary steps—internal controls, independent auditors, a board-level committee—sound like real oversight. But they are voluntary. Companies can opt out, and the accord explicitly leaves room for future regulation while keeping everything self-policing. That is the problem. The biggest firms get to set their own rules and police themselves, and the public gets to trust them.
The public should watch whether the companies actually follow through on the voluntary steps. The paper would prefer a narrow rule requiring companies to disclose safety failures they hid, rather than a voluntary system that rewards secrecy. The accord is not meaningless, but it is a weak foundation for managing real risks.
Source material: “Trump says top tech firms have signed accord to 'self-police' AI development,” NPR.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.