US President Donald Trump and tech executives have signed a voluntary agreement aimed at ensuring their AI models “behave as intended”.
The deal comes amid growing concerns over the security of increasingly capable frontier AI systems. Trump countered the slowdown by citing competition with China. The agreement places much of the responsibility of AI risk management on companies developing systems.
The White House agreement, called the “Joint Commitment to Border Responsibilities,” calls for internal controls, independent external audits and board-level oversight “to ensure its models do not hack or access technical systems in unintended ways.” The agreement also leaves open the possibility of turning those commitments into laws or regulations, saying it may make sense to do so over time.
“I think I’m seeing tremendous self-policing. And they understand that they have to self-police,” Trump told reporters at the White House.
Google CEO Sundar Pichai, Anthropic CEO Dario Amody, Meta CEO Mark Zuckerberg, OpenAI President Greg Brockman, Nvidia CEO Jensen Huang and XAI CEO Elon Musk signed the agreement at the White House.
The meeting discussed AI security, industrial growth and data center development, while Trump ruled out a US-China joint venture to develop AI.
Trump came on the same day signed An executive order directing the executive branch to use “superintelligence” instead of “artificial intelligence,” “more aptly captures the promise, potential, and rapidly advancing capabilities of this technology.”
Related: Trump says US will form ‘AI force’ and appoint AI czar: report
