The most powerful people in tech walked out of a White House summit with US President Donald Trump on Tuesday having promised to keep their own creations in check.
Trump said the executives, among them Nvidia’s Jensen Huang, OpenAI’s Greg Brockman and Anthropic’s Dario Amodei, had signed a “morally binding” commitment to build safeguards into artificial intelligence (AI).
The pact carries no legal force and sets out no clear penalties for breaking it.
“They’re going to police themselves,” Trump told reporters after the meeting.
The lunch brought together 31 guests, including Meta’s Mark Zuckerberg, Elon Musk, Amazon founder Jeff Bezos, Microsoft’s Satya Nadella and Google’s Sundar Pichai and comes as Washington’s mood on AI continues to sour.
Lawmakers face growing pressure from voters to curb or slow the technology’s development over fears it could displace workers and upend society. A run of incidents over the past months in which OpenAI’s autonomous systems slipped out of their constraints has added to those fears.
The first was a breach of Hugging Face, a widely used platform for sharing AI models, in July. On Monday, OpenAI cancelled the release of its latest model, Astra 6.1, after internal testing found it did not meet safety standards.
Here are the key takeaways from the summit.
Tech bosses ‘self-police’
The document, dubbed the White House Accord, sets out the industry’s commitment to regulate itself. It is not legally enforceable, but it leaves the door open “to codify these steps into laws or regulations” if that makes sense over time.
Zuckerberg said the accord aims to assure customers that AI “works in the way that we intend.” He said it includes “robust internal controls” and several layers of auditing, from internal risk reviews to external auditors and oversight by company boards, and called it a starting point for industry-wide standards.
Musk put it more bluntly. He said the executives had essentially agreed to “grading each other’s homework, which is a lot better than if people just grade their own homework for obvious reasons.”
Trump said he would name someone to oversee the agreement in the coming days and also suggested a committee of about 10 people to “watch over the whole enterprise”.
Amodei’s firm has clashed bitterly with the White House over AI safety, and he struck a more cautious note than his peers.
“As the president has said, whoever wins AI wins. I think that’s very important,” Amodei said after the meeting. “But I think the technology has very real risks, and the mechanism, how we address those risks, is still under discussion.”
His company recently warned investors of potential existential risks to humanity in the prospectus for its planned stock market debut. Anthropic is aiming to go public as early as November.
Sweeteners for data centres
Advised on AI expansion by Huang and his former AI czar David Sacks, Trump has pushed for unfettered development of the technology to outpace China and to fuel a US economy that increasingly relies on AI-driven growth.
That push runs through data centres, the vast energy-guzzling computing facilities AI depends on. In both parties, they have become a flashpoint in midterm election races, as communities worry about rising electricity bills and pressure on water and land.
An NPR/PBS News/Marist poll carried out on 14 and 15 September found that 65% of registered voters opposed data centres being built in their own communities, with 43% strongly opposed.
More than 80% of Democrats were against them, while Republicans were split, with 44% opposed and 46% in favour.
On 16 September, the House of Representatives voted 417 to 3 to pass the Ratepayer Protection Act. The bill requires states to consider rules that make data centres pay for grid upgrades and rising electricity costs instead of passing them on to households. It has yet to clear the Senate.
Earlier this month, Trump wrote on Truth Social that communities turning away data centres were choosing to be “backwards and poor”. After Tuesday’s meeting, he struck a softer tone, saying AI companies “will be very generous” with communities that welcome them.
“They’ll help with their schools, they’ll help with their teachers,” he said.
Goodbye artificial intelligence
Trump also confirmed he would officially rename the technology “super intelligence” instead of artificial intelligence “because it’s not artificial.”
Hours later, he signed an executive order telling the federal government to use the terms super intelligence and SI whenever possible.
“‘Super Intelligence’ more appropriately captures the promise, potential and rapidly advancing capabilities of these technologies,” the order said.
How Brussels does it
Across the Atlantic, the EU has written its safeguards into law. The AI Act, which Brussels bills as the world’s first comprehensive law on AI, applies to any company selling AI in the bloc, wherever it is based.
Since August 2025, the companies behind general-purpose AI models, such as those built by OpenAI, Anthropic and Google, have been legally required to document how their models work, share key information with businesses that build on them, respect EU copyright law and publish a summary of the data used to train them.
The most powerful models face stricter rules. Their makers must test them for risks, including through adversarial “red-teaming,” report serious incidents to Brussels and keep their systems secure from cyberattacks.
These rules have actual teeth. Since 2 August, the European Commission’s AI Office has been able to demand information from companies, run its own evaluations of their models and order them to restrict or withdraw a model from the EU market.
Companies that break the rules face fines of up to €15 million or 3% of their global annual turnover, whichever is higher. For Big Tech, that could mean billions.
A voluntary code of practice sits alongside the law to help companies comply. But unlike the White House pact, signing it is optional — obeying the AI Act is not.