President Donald Trump mentioned that tech leaders he met with on the White Home on Tuesday had signed a “structure” to police themselves amid dire warnings from synthetic intelligence leaders in regards to the know-how.
Home Speaker Mike Johnson mentioned the accord is “voluntary” for the AI firms and Trump harassed that the trade will probably be “self-policing.” When requested whether or not the settlement is a binding one, Trump mentioned it is “morally” binding.
The assembly got here as AI leaders are issuing warnings about the specter of AI and calling for steps to gradual it down earlier than it outsmarts people. Following these warnings, lawmakers are weighing what — if any — motion ought to be taken.
President Donald Trump speaks to the press after a gathering with know-how executives about synthetic intelligence on the White Home in Washington, September 29, 2026.
Kent Nishimura/AFP by way of Getty Photos
Together with Johnson and members of Trump’s Cupboard, Meta CEO Mark Zuckerberg attended the assembly together with Elon Musk, Anthropic CEO Dario Amodei, Google CEO Sundar Pichai, Jensen Huang from Nvidia, Palintir’s Alex Karp and Shyam Sankar, Amazon’s Jeff Bezos, and Greg Brockman from OpenAI.
Johnson mentioned the tech leaders signed on to a doc known as “The White Home Accord on Superintelligence: A Joint Dedication on Frontier SI Obligations.”
“It is a assertion of ideas that you’re going to see, a press release of requirements, commitments which might be voluntary on behalf of the trade,” Johnson mentioned. “It units forth expectations. It is in brief abstract strong inside controls and layers of inside and exterior overview. That is vital since you want individuals to be offering oversight and transparency, and that is what these leaders are going to do,” Johnson mentioned.
The speaker mentioned Congress would “proceed to information the trade’s growth in a secure method,” however stopped in need of asserting any motion by Congress to implement these ideas.
President Donald Trump listens as Elon Musk speaks to reporters on the White Home in Washington, Sept. 29, 2026, after assembly with high executives of AI companies.
Kevin Dietsch/Getty Photos
Trump indicated there could be a “committee” to supervise AI, although it is unclear who can be on the committee or whether or not its selections can be binding.
“We’re additionally occupied with forming a committee of kinds, the place we put perhaps 10 individuals on that committee,” he mentioned. “It could possibly be from that group, in order that we are able to so the committee can watch over the entire enterprise.”
ABC Information has requested the White Home for an official of the copy of the accord.
Trump later shared a screenshot of the accord, although it’s skinny on particulars. It says the signatories “imagine” that every AI firm “ought to implement” a collection of controls and audits.
OpenAI and Anthropic have already not too long ago dedicated to implementing third-party evaluators and it isn’t instantly clear if the settlement instructs them to do something totally different.
The accord within the screenshot additionally emphasizes every firm’s particular person accountability to make sure that its know-how “builds belief with prospects and the general public.”
It acknowledges that the steps outlined usually are not codified into formal legal guidelines or rules, however leaves open that risk for the longer term.
President Donald Trump speaks with Meta CEO Mark Zuckerberg on the White Home driveway following a luncheon for tech leaders within the East Room, in Washington, September 29, 2026.
Kevin Lamarque/Reuters
Zuckerberg mentioned the settlement was “a begin and an accord that the entire trade might come to.”
He mentioned it was a set of “ideas and commitments” round “constructing a number of layers of auditing and controls” and that firms’ boards would independently overview stories from auditors.
“I believe that these are good steps ahead,” he mentioned.
When requested by a reporter why there should not be extra stringent guardrails than the voluntary doc, Trump responded that the AI CEOs “love” America. However when pressed by a reporter about whether or not it requires greater than “love,” the president disagreed.
“No, it requires this,” he mentioned pointing to his head, “and so they’re the neatest individuals on the planet.”
Trump has mentioned he opposed guardrails and rules on the know-how or something that will “stifle” its progress. Simply final week whereas addressing the United Nations Common Meeting, Trump mentioned the U.S. “rejects” a world try to manage AI.
President Donald J. Trump and Speaker of the Home Mike Johnson hear as CEO and co-founder of Anthropic Dario Amodei responds to a query from the information media outdoors the West Wing of the White Home in Washington, September 29, 2026.
Shawn Thew/EPA/Shutterstock
Tuesday’s assembly wasn’t the primary time that high tech leaders have met with Trump on the problem. The Meta CEO additionally attended the state dinner on the White Home for Chinese language President Xi Jinping final week, together with the CEOs of OpenAI, Nvidia, Microsoft, Google, AMD, SpaceX and executives from Amazon and Apple. Amodei additionally had dinner with Trump late on Sunday evening.’
Johnson advised reporters two weeks in the past in regards to the assembly, saying there can be “a deliberate dialogue in regards to the accountability of the businesses to take care of security, and what position, if any, the federal government has to play in that.”
Whereas Johnson known as on the businesses to “self-regulate,” he additionally additionally harassed that the U.S. shouldn’t lose its aggressive edge within the house.
“We can’t have a moratorium on the event of AI, as a result of then we are going to lose our edge to China, and that has severe nationwide safety implications for,” Johnson mentioned.
On Monday, Johnson mentioned the the fears of the specter of AI is a Chinese language psychological operation and are politically motivated.
“That is what Democrats do. They glom on to one thing and say, ‘That is doom and gloom and everybody will die if we do not win,'” he advised Fox Enterprise. “Do not buy it. It’s a Chinese language psyop.”
ABC Information’ Emily Chang contributed to this report.
