US President Donald Trump privately met OpenAI chief executive Sam Altman last week, days before the two publicly diverged over whether the development of advanced artificial intelligence should slow down, according to MS NOW.
The US news outlet, citing three people familiar with the matter, reported that the meeting took place backstage at a Republican voter conference on Thursday. One source said Altman requested the discussion, which focused largely on AI and its growing capabilities.
Few details of the conversation were disclosed. It followed public accusations by a former OpenAI researcher that the company and rival Anthropic were acting irresponsibly and gambling with people’s lives.
In the days after the meeting, Altman and Trump outlined sharply different approaches to the technology.
Altman wrote on Sunday evening that slowing development would carry a significant cost, but argued that competitive pressure on the United States could not justify recklessness or allow AI capabilities to outpace safety measures and oversight.
According to the report, Altman also said over the weekend that OpenAI would postpone plans for an initial public offering this year because of safety concerns.
In an unusual convergence, Altman, Elon Musk and Anthropic chief executive Dario Amodei all argued over the weekend for slowing AI development before the technology advances beyond humanity’s ability to control it.
Trump pushed back on Monday, writing on Truth Social that AI needed only the leadership of a strong, intelligent president.
In a separate post, he warned that regulation would drive the industry towards economic ruin and described AI as an unprecedented engine of economic development.
The disagreement highlights competing pressures over how quickly the United States should pursue more powerful AI systems while addressing concerns about their safety.
Trump has repeatedly argued that regulation should not obstruct innovation. Shortly after returning to office in January 2025, he revoked a Biden-era directive establishing safeguards for the technology.
According to the report, he signed another executive order in December aimed at establishing a national AI framework and overriding state rules, arguing that American developers needed freedom from burdensome regulation.
In June, however, Trump signed a separate order directing federal officials to work with AI companies on a voluntary framework that would give the government access to new models for up to 30 days before their release.
Altman supported that initiative, saying the United States should continue developing leading AI models while ensuring they were safe.
A White House official told MS NOW that the framework had been completed by an August 1 deadline and that the administration met AI companies the following day to discuss it. A person familiar with that meeting confirmed Anthropic attended.
The final framework had not been publicly released, according to the report.
Trump and Altman were last seen together publicly in June, sitting beside one another at a G7 working lunch attended by world leaders and technology executives.
The White House and OpenAI did not immediately respond to MS NOW’s requests for comment.
