US President Donald Trump announced on Tuesday that he had signed a voluntary agreement with a group of leading artificial intelligence company executives requiring internal and external reviews of their systems. The announcement came during a White House meeting convened to discuss growing concerns among Americans about the technology’s risks and the spread of its infrastructure across the country.
The meeting brought together Trump and House Speaker Mike Johnson with senior executives from technology companies spending hundreds of billions of dollars on the infrastructure needed for artificial intelligence. The investments include data centres built in American communities, cities and towns, and have sparked public opposition among supporters of both the Democratic and Republican parties.
The agreement sets out four voluntary steps for reviewing systems
The group included chief executives who had previously called for a slowdown in artificial intelligence development, following prominent incidents in which agents powered by the technology went out of control. Trump published the agreement on his Truth Social account, setting out a voluntary framework that could eventually pave the way for regulation of the sector through laws and regulations. The agreement focuses on four steps companies will take voluntarily.
It calls for the implementation of “strong internal controls” and the use of an “independent external auditor” to assess how effective those controls are, as well as the creation of a committee linked to company boards that would review reports submitted by internal and external auditors. The agreement also proposes that these measures could, over time, be codified in formal laws and regulations.
The meeting also featured warnings about the agreement’s limits and how it would be implemented. Dario Amodei, chief executive of Anthropic and one of the leading advocates of slowing artificial intelligence development, said: “This technology carries very real risks.” He said the mechanism to be adopted and how those risks would be addressed were still under discussion.
The agreement was signed by Trump and Amodei, as well as Google chief executive Sundar Pichai, Meta chief executive Mark Zuckerberg, OpenAI president Greg Brockman, Nvidia chief executive Jensen Huang, and Elon Musk as the founder of xAI, which has become part of SpaceX.
Trump links ethical commitment to government oversight
Trump said he would select about 10 people to serve on a committee that would “oversee the entire project”, describing the agreement as “morally binding”. He added that in the coming days, after consulting the industry, he would appoint someone to oversee its implementation. The agreement was signed after Trump pledged earlier on Tuesday that the US government would not restrict artificial intelligence development or support calls for “controls” or other restrictions that would hinder its growth.
He made the comments while launching the AI-powered government chatbot, designed to help Americans use government services. During the launch, Trump said: “I will not stand in the way of the growth of a technology that will surpass the Industrial Revolution in importance.
We cannot stop it, and we are far ahead in it”. The chatbot was unveiled during an all-day event in Washington, where several administration officials spoke of a “golden age” led by technology that Trump insists on calling “super intelligence” rather than artificial intelligence.
Government chatbot expands online document services
Secretary of State Marco Rubio said the chatbot available through the government website would soon allow people applying for a passport for the first time to complete the process entirely online, without having to visit a processing centre in person — a step currently required under existing procedures.
Regarding data centres, Trump said the chief executives had agreed to take steps to contain public anger and bipartisan opposition to the widespread construction of these facilities in cities and towns. The measures include providing additional financial support to local schools or taking steps to help reduce energy costs.
Trump presented the race to develop artificial intelligence as a direct global competition between the United States and China, insisting that his country was ahead in the field. He said the race would end with a winner and a loser, and that there would be “no second place”.
The Trump administration is facing mounting pressure to address the risks artificial intelligence technologies could pose to the public, after OpenAI and Anthropic reported that programmes operated by them had gone out of control and breached customers’ systems.
The breaches prompted researchers to warn that some technology developers believe it could kill humans within a decade, as lawmakers from both parties increasingly call for action to curb the risks of the emerging technology.
