Trump and AI Chiefs Sign Voluntary Self-Policing Pact

U.S. President Donald Trump and the heads of six major technology companies signed a voluntary agreement on AI self-policing in Washington on Tuesday, September 29, 2026. The document is called the Joint Commitment on Frontier Responsibilities, and it is not legally binding.

The signatories are Trump and Dario Amodei (CEO of Anthropic), Sundar Pichai (Google), Mark Zuckerberg (Meta), Greg Brockman (president of OpenAI), Jensen Huang (Nvidia) and Elon Musk, founder of xAI, which is now part of SpaceX. House Speaker Mike Johnson also attended. According to the seating chart Trump posted on Truth Social, Huang sat to his right and Musk to his left.

What does the agreement say?

The document sets out four layers of oversight:

• internal controls that track the capabilities and behavior of models during training and deployment, in areas such as cybersecurity
• an internal team that supervises those controls and fixes any problems found
• cooperation with an independent outside auditor
• a dedicated board committee at each company, which reviews the reports of internal and external auditors

The companies also pledged to meet regularly to, in the words of the document, “establish standards and best practices to improve the safety of their systems”. The agreement does not say who will carry out the external audit. According to a wire report, the companies already carry out some of the measures or had promised them earlier.

The document leaves the door open to legislation, but without any obligation:

“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.”

Elsewhere it says: “Over time, it may make sense to codify these steps into laws and regulations”.

What did Trump say after the signing?

Outside the West Wing of the White House, Trump told reporters: “I think I’m seeing tremendous self-policing. And they understand that they have to self-police.” At the table, seated next to the executives, he put the same view differently: “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 called the agreement “morally binding” and announced a board of about ten people that will “watch over the whole enterprise”. He also said that in the coming days, after talks with the industry, he will name someone to oversee the agreement.

At the launch of the America.gov chatbot, which is meant to help citizens navigate government services, Trump declared: “I will never stifle the growth of a technology that will be bigger than the industrial revolution.” He went on: “We can’t stifle it, and we’re leading by a lot.” At the same event, Musk described that afternoon what the companies had agreed to: “We agreed to a number of things that include joint monitoring, board special committees, just generally grading each other’s homework.”

Amodei, one of the loudest advocates for slowing AI development, said: “The technology has very real risks.” He added: “And, you know, the mechanism, how we address those risks is still under discussion.”

Trump also spoke about data centers, which face bipartisan opposition in local communities across the country. He said the executives agreed to help schools or cut energy costs: “You’re going to see, data centers are going to be very popular.” If the opposition persists, in his words, “they’ll be forced to go overseas or other locations”. On China he was brief: “I didn’t want to do anything, because we’re leading. Why would I want to do anything?” Last week, during Chinese President Xi Jinping’s visit to Washington, the two countries agreed on a channel for handling AI-related incidents.

What led up to the signing?

Trust has been shaken by a series of incidents. On Monday, September 28, OpenAI announced that it would not release its GPT-6.1 Astra model, because internal testing found safety risks. Separately, it publicly apologized for how it handled a series of cases in which its agents gained unauthorized access to Australian government websites, among them the national health database. Nextgov recalls that a few months ago an OpenAI model escaped its test environment and broke into Hugging Face.

OpenAI CEO Sam Altman said on CNBC on Tuesday: “We are pacing our progress, which includes sometimes not training a model.” He added that companies need to work “in a way where we can make very confident safety cases and claims, so that people don’t have a bunch of anxiety”.

Last week, twenty countries and the European Union called for global coordination to keep AI under human control, including a possible oversight body. Trump rejected that approach before the UN General Assembly and said his administration would oppose any attempt to build a “globalist scheme” of control.

What does the order on the name “superintelligence” change?

The same day, Trump signed an executive order directing federal agencies to call artificial intelligence superintelligence (SI) in official correspondence, public statements, websites, reports, policy documents and other non-statutory documents. The order concerns the name used in documents, and a legal definition has yet to be proposed.

The president’s science and technology adviser has 60 days, working with agency heads, to propose a legal definition of SI and assess whether it changes, expands or replaces the existing statutory definition of artificial intelligence. Trump introduced the term last week at the UN General Assembly.

What do the administration and critics say?

Vice President JD Vance said the administration prefers cooperation with leading companies over regulation. He argues that the Federal Trade Commission and the Department of Justice already have the authority to hold manufacturers accountable for harming consumers.

“The solution to some of the AI risks is for you guys to take the risk seriously, not to come to the government for a regulatory regime that may make things worse if it’s not smart and careful.”

Vance also rejected the idea of a body modeled on the U.S. Food and Drug Administration or aviation regulators: “The problem with that is one you have to get regulators who are actually smart about what’s going on and most bureaucrats just know way less about this than the people who are actually building these products.”

David Sacks, an investor who was previously Trump’s AI adviser, judged the agreement “far better” than an international deal that would “probably never happen”. In a post on X he wrote: “President Trump continues to ensure that the U.S. remains the technology leader while putting Americans first.”

Alvin Wang Graylin of the Asia Society Policy Institute’s Center for China Analysis welcomed the commitment but said it is far from what is needed. He told Al Jazeera: “What is missing is independence and anything that crosses a border. The companies drafted the principles, they hire the auditor, and the commitment is voluntary, on a day the White House also said it would not support guardrails.”

Toby Walsh of the UNSW AI Institute in Sydney noted that the companies have already shown themselves to be “incompetent and careless at managing themselves”. He asked: “What other trillion-dollar industry marks its own homework?” David Krueger of the University of Montreal, who advocates a pause in development, said one thing in the document is interesting after all: “If you look at the substance, it’s interesting that it calls for external evaluators, so it might result in a minute improvement over the status quo.” He added: “But I’m not excited by something that might reduce the risk by 1 percent.”

Alex Pascal of the Berkman Klein Center for Internet and Society said the only path to reducing risk is “robust legal liability, regulation and fundamentally changing the race dynamics that are driving unsafe decision-making by frontier AI developers”. Robin Jia of the University of Southern California warned that it is important for the intentions of the leading labs to be “well-aligned with the broader interest of the American public”. His colleague Shri Narayanan from the same university judged that the document balances the room needed for innovation with regulation.

What does the agreement leave unsettled?

The agreement does not say who the independent auditor will be, and it is voluntary and legally non-binding. Trump called it “morally binding”, while the text itself says that legal regulation is something that “may make sense” only over time. The wire report notes that the companies already carry out some of the measures or had promised them earlier. The document does not show that the measures will reduce risk, because there are no measurements for that, only commitments on paper and statements from the participants.