President Trump summoned the world’s most powerful technology executives to the White House on Tuesday, and praised the assembled billionaires as “very brilliant” minds who wanted to make the world a better place through job creation, the construction of data centers and the rapid expansion of artificial intelligence.
As president, Mr. Trump faces questions about whether A.I. should be regulated, and particularly whether restrictions should be put in place to protect humans from technology that has learned to outmaneuver existing safeguards.
But in the end, convening all of those billions and all of that brain power did not result in a commitment to formal regulation. Instead, they signed a voluntary agreement to adopt safety policies that include risk reviews, auditing and evaluations of systems by third parties. They also agreed to adopt Mr. Trump’s preferred rebranding of A.I.: “super intelligence.”
“This is going to be all for the good, and we’re going to have it be nice and safe,” Mr. Trump said on Tuesday afternoon, when he and the tech leaders took questions outside the White House. “And they’re really going to be policing each other, and that’s the way it works.”
Asked if the agreement would be binding, Mr. Trump said, “I think it’s morally binding.”
Inside the White House luncheon, Mr. Trump’s seating chart (which he posted on social media) revealed a powerful assemblage of executives who are deeply divided on the need for A.I. safeguards. The meeting also included top officials, such as Speaker Mike Johnson and Vice President JD Vance, who believe that formalized safeguards would hurt American competitiveness.
In the middle of it all was Mr. Trump, who has punted to a group of executives to decide how best to wrangle the rapidly evolving technology, or whether it should be wrangled at all.
“There’s going to be a tremendous self-policing aspect,” Mr. Trump said.
The one-page document is short on details and relies on the companies to follow through with their promises. It is not an official order, and it is unclear who would enforce the guidance. It was signed by Mr. Trump and the leaders of Anthropic, Google, Meta, Nvidia, OpenAI and SpaceXAI.
The agreement echoes voluntary safety commitments by A.I. companies made during the Biden administration. The companies agreed to test their systems for safety and to report safety and security incidents. But the commitments were nonbinding and largely lost at the start of the Trump administration, when the president rolled back safety requirements.
The business leaders invited to the White House on Tuesday included Dario Amodei, the chief executive of Anthropic, who has called for federal safety regulations, including the testing of models. Earlier this month, Mr. Amodei called for an industrywide slowdown of A.I. development, citing safety concerns.
The company has also been embroiled in a legal fight with the administration, and has been barred from working with the Pentagon. The two sides disagreed in negotiations over a $200 million contract, with the company’s insisting that its A.I. technology not be used in mass surveillance of Americans or for autonomous lethal weapons. Speaking to reporters on Tuesday, Mr. Trump called Mr. Amodei “fantastic.”
Also on the guest list was Jensen Huang, the chief executive of Nvidia and a close Trump ally, who has called the safety concerns exaggerated. Mr. Huang has implored the administration and lawmakers to keep a hands-off approach to regulating the industry, saying rules would allow rival China to speed past the United States on A.I. leadership.
When the group appeared with reporters, the president called Mr. Huang a “very handsome guy” and asked him to speak.
“We’re going to advance this technology that is so consequential to humanity, so consequential to the future prosperity of our nation,” Mr. Huang told reporters. “We’re going to do it responsibly and safely.”
Others invited to the luncheon included Jeff Bezos, the founder of Amazon; Greg Brockman, the president of OpenAI; Alex Karp, the chief executive of Palantir; Sundar Pichai of Google; Mark Zuckerberg of Meta; and Elon Musk of SpaceXAI. The White House chief of staff, Susie Wiles, and Treasury Secretary Scott Bessent were there. David Sacks, Mr. Trump’s former top adviser on A.I. and a venture capitalist, also attended.
Before the luncheon, Mr. Trump told the audience at an event unveiling America.gov, an A.I.-powered search tool, that the U.S. government was championing the technology in an effort to counter China, a major rival in the race to develop A.I.
“We don’t want to stifle growth,” the president said.
Mr. Trump, who hosted a state visit last week for Xi Jinping, the Chinese leader, said that the two did not spend much time discussing opportunities for partnerships or issues related to competition. Mr. Xi said in remarks at the White House that both nations shared the responsibility to develop A.I. with “the well-being of the people” in mind.
But in front of a crowd of executives and cabinet members, who were gathered in an auditorium for a presentation that took on the veneer of an Apple product-launch event, the president reiterated a winner-take-all mentality.
“You’re going to have a winner, or you’re going to have a loser,” Mr. Trump said.
Mr. Trump has argued that American dominance over A.I. technology is imperative for national security and economic growth.
With his embrace of A.I. and support for the development of data centers throughout the country, Mr. Trump is also increasingly out of step with the public. Recent polling shows that a majority of Americans are growing more concerned about A.I. and want the federal government to do more to regulate the technology.
On Tuesday, Mr. Trump again came down on the side of data centers, calling concerns about them a Democratic hoax. Flanked by executives at the luncheon, Mr. Trump predicted that the public would eventually embrace the facilities.
“You’re going to see data centers are going to be very popular, because this is the group, and they’re the ones that want to, and they’re also the ones that want to see communities that are safe and happy,” Mr. Trump said. “Otherwise, they’ll be forced to go overseas or other locations.”
As the public has soured on A.I. and the prospect of data centers mushrooming around the country, the debate over technology safety has intensified in recent weeks. Several rogue A.I. clients and models have worked together to breach websites, steal data and then cover their tracks, all while appearing to show an understanding that what they were doing was wrong. Senior leaders in Congress have acknowledged the need for A.I. safety guardrails.
In April, Anthropic announced it was holding back the release of one of its models, Mythos, because it was too dangerous for the public. In July, OpenAI models broke out of a restricted testing environment and hacked the A.I. start-up Hugging Face. In early September, leaders of Anthropic, OpenAI and Google DeepMind called for slowing development of advanced A.I. models.
Last week, an OpenAI model was found to have meddled with U.S. government websites including the Education Department, the Securities and Exchange Commission and Commerce Department’s census site. And this week, OpenAI said it would not release its latest model, GPT-6.1 Astra, over safety concerns.
Within the White House, senior leaders are at odds on how to regulate the industry. Mr. Bessent has floated the idea of an open dialogue with China on safety standards. He has also stressed that A.I. companies must be held accountable for harms created by their models.
As his day with the billionaires wrapped up on Tuesday, Mr. Trump said that he was considering a committee of about 10 people focused on A.I. safety, and that he would name an A.I. czar in the coming days.
He ended his final appearance with reporters alongside Mr. Musk, another attendee who has worked himself back into Mr. Trump’s good graces after a spat last year over domestic policy legislation. He was seated near Mr. Xi and Mr. Trump at last week’s state dinner, and was at Mr. Trump’s side at the luncheon on Tuesday.
Mr. Xi, the president said, “will take every data center that we don’t want.”
Standing nearby, Mr. Musk agreed.
“He would love that, every single one, because you know why? They’re smart,” Mr. Trump continued. “They want every data center. They want it because that’s great for China. Well, it’s great for us, and we’re leading by a lot, and we’re going to keep it that way.”
The post At White House, Trump Asks A.I. Titans to Police Themselves appeared first on New York Times.
