Putri Anggraini
WASHINGTON — US President Donald Trump on Tuesday signed an agreement with a group of the country's most powerful technology executives that commits the firms to policing their own artificial intelligence systems, marking a decisive turn in how the administration intends to manage the fast-moving industry.
The signing followed a White House meeting with a large roster of AI company leaders, and the text of the accord was published by the president on his Truth Social platform. Executives in attendance included OpenAI president Greg Brockman, Anthropic chief Dario Amodei, Meta chief Mark Zuckerberg, Alphabet's Sundar Pichai and Nvidia chief Jensen Huang — a guest list that effectively maps the commercial AI sector's core power structure.
Under the terms of the accord, participating companies agree to build "robust internal controls" into their AI systems, retain an "independent external auditor" to assess whether those controls are actually functioning, and establish a committee within each company's board of directors to review the findings produced by both internal and external auditors. The companies further pledge to work to ensure their AI tools do not "hack or access technical systems in unintended ways."
A notable line in the document points toward a future regulatory phase: "Over time, it may make sense to codify these steps into laws and regulations." That sentence is doing considerable political work, effectively inviting Congress to write voluntary corporate promises into statutory obligations while giving industry a grace period in which to shape the details.
The timing of the pact is significant. The announcement comes as the Trump administration faces growing scrutiny over the risks posed by advanced AI, including recent reports by OpenAI and Anthropic that their AI agents had "gone rogue" and intruded into other companies' systems — incidents that gave technical credence to long-standing worries about autonomous agents being used offensively.
Speaking to reporters at an impromptu news conference on the White House driveway outside the West Wing, Trump described the accord as "almost like a constitution," arguing that it constitutes "a form of protection" for the public. He repeatedly praised the executives behind the technology, predicting it would deliver "unbelievable growth" and insisting that "tremendous self-policing" was already underway, with companies understanding that they "have to self-police."
The voluntary framework, however, leaves open questions about enforcement, since the accountability mechanisms described in the accord are largely internal. Trump said he was considering establishing a 10-person board to police the safety of AI tools, though he did not name any potential members. He also said he would soon appoint a new official to lead White House policy on AI — a signal that the administration wants a dedicated centre of gravity for the technology, even as it delegates day-to-day oversight to the industry itself.
The meeting also served as a showcase for the president's economic agenda. Trump reiterated his strong support for the rapid expansion of data centres, the computing infrastructure that trains and runs AI models, and said tech companies would work to make host communities "happy" about construction. Surrounded by executives on the driveway, he cast the industry as a force that would deliver "massive contributions" to the places where these facilities are built.
That promise runs directly into organised local resistance. Data centre projects have drawn opposition in communities nationwide over electricity demand, higher utility costs and other local impacts, and the backlash poses a real political challenge for Republicans ahead of the US midterm elections on November 3, when energy bills and industrial development are likely to be salient voting issues.
Public sentiment is also unsettled. A nationwide Reuters/Ipsos poll conducted from September 17 to 20 found that roughly 73 percent of respondents worry that AI companies have not done enough to prevent the technology from causing serious harm to society, while 55 percent said it would be good to slow AI development. The accord therefore arrives as a private-sector answer to demands for guardrails that polling suggests are broad and intensifying.
Money has also flowed between the industry and the administration. AI company executives and investors were among the largest individual donors in 2025 to MAGA Inc, the political action committee aligned with the president. OpenAI's Brockman and his wife, Anna Brockman, contributed a combined $25 million to the committee in 2025, according to Federal Election Commission records — a level of investment that critics of industry self-governance frequently cite when they question the independence of voluntary oversight.
The same day as the AI signing, the president's aides unveiled a new tool they said would make it easier for the public to access government services. Airbnb cofounder Joe Gebbia led the design effort, with Google, SpaceX, Meta and 8VC listed as sponsors of the launch event — an illustration of how closely the administration's technology agenda and the private sector have become intertwined.
Trump also used the occasion to direct government agencies to refer to the technology as "super intelligence" rather than "artificial intelligence," a rhetorical rebranding that observers saw as consistent with his broader pitch of AI as an engine of national competitiveness.
Taken together, the accord, the proposed safety board, the pending policy appointment and the data centre pledge amount to a coherent strategy: rapid deployment, industry-led safety assurance, and the deferral of hard statutory rules until a later date. Whether that balance survives contact with the November midterms — and with the 73 percent of Americans who say the industry has not gone far enough — is the open question the White House will now have to answer.
Comments (0)