US President Donald Trump and the leaders of six major American technology companies have signed a White House agreement introducing tougher safeguards aimed at preventing increasingly powerful artificial intelligence systems from behaving in unintended or dangerous ways.
The agreement, known as the White House Accord on Super Intelligence – Joint Commitment on Frontier Responsibilities, was signed on Tuesday by Trump alongside Google chief Sundar Pichai, Anthropic CEO Dario Amodei, Meta CEO Mark Zuckerberg, OpenAI president Greg Brockman, xAI founder Elon Musk and Nvidia CEO Jensen Huang.
Under the accord, companies developing advanced AI models have committed to strengthening their internal safety systems, monitoring how their technology behaves and responding rapidly when problems are detected.
A central requirement is that AI systems should be prevented from hacking or gaining unintended access to computer networks and other technical infrastructure. Each participating company will establish an internal team responsible for checking that monitoring, detection and control mechanisms are functioning properly.
The companies have also agreed to submit their safeguards to independent external audits and establish independent board-level oversight to assess whether the commitments are being followed.
The agreement comes amid growing concern over how quickly advanced AI systems are developing, including fears that future models could improve their own capabilities faster than developers can adequately understand or control them.
Zuckerberg said the initiative was intended to give both customers and the wider American public greater confidence that advanced technology would operate in the manner its developers intended.
“The basic idea is that we want to give the American people and our customers confidence that the technology works in a way that we intend,” he said.
Zuckerberg described the agreement as an initial industry-wide step rather than the final word on AI safety.
“The idea isn’t that this is the only thing that we will ever do. This is a start and an accord that the whole industry can come to,” he added.
Trump also used the meeting to sign an executive order directing the federal government to adopt the terms “super intelligence” and “SI” in place of “artificial intelligence” and “AI.”
The White House argued that “Super Intelligence” better reflects the capabilities emerging from the latest generation of technology, rather than suggesting that such systems merely imitate or automate human intelligence.
The agreement follows renewed debate within the technology industry over whether development of increasingly capable models should be deliberately slowed to allow safety mechanisms to catch up.
Anthropic’s Dario Amodei recently called for a more measured pace of development, arguing that researchers need sufficient time to establish effective controls. Musk and OpenAI CEO Sam Altman subsequently backed the call for a more balanced development rate.
Concerns have also intensified following reported incidents involving advanced AI agents obtaining unauthorised access to external computer systems, including an episode in which OpenAI agents reportedly escaped a testing environment and accessed servers belonging to AI platform Hugging Face.
Trump, however, has rejected proposals for the United States to deliberately slow the development of advanced AI. Earlier this month he dismissed warnings about runaway technology as a “hoax” and insisted his administration would not obstruct the industry’s expansion.
“We will not in any way hinder or stifle the Growth of this incredible Industry,” Trump wrote.
Nvidia chief Jensen Huang has similarly supported stronger safeguards while opposing attempts to deliberately restrain technological progress. He has argued that development should proceed rapidly, provided companies do not release systems before they are ready or adequately safe.
The White House agreement also follows last week’s meeting between Trump and Chinese President Xi Jinping, after which Washington and Beijing agreed to establish a dedicated hotline for serious AI-related incidents.
China has likewise called for stronger oversight of rapidly advancing systems, arguing that regulation and governance need to develop alongside the technology to prevent it from getting out of control. Beijing has nevertheless resisted calls for an international slowdown, while warning against both alarmism and an escalating technological race between major powers.
