President Donald Trump signed a new executive order on Tuesday that asks AI companies to voluntarily submit their most powerful models to the government for review before public release. The order represents a significant retreat from earlier proposals after sustained pressure from Silicon Valley.
The final version gives companies a 30-day window for optional government testing and evaluation of new AI models. This marks a substantial compromise from the original 90-day mandatory review period that was proposed in earlier drafts. Industry leaders had pushed for an even shorter timeframe of around two weeks, arguing that longer delays would hurt American competitiveness against China.
This development highlights the ongoing tension between AI safety advocates who want stronger oversight and tech companies pushing for minimal regulation. The compromise reflects Trump’s broader approach of trying to balance safety concerns with maintaining America’s lead in AI development. The administration has repeatedly emphasized that overly restrictive policies could hand advantages to international competitors, particularly China.
The order was originally scheduled for signing in late May with top Silicon Valley CEOs in attendance. However, Trump delayed the ceremony after receiving pushback from influential figures including venture capitalist David Sacks, who previously served as the White House AI czar. Instead, Trump signed the revised order privately on Tuesday.
The final text explicitly states that nothing in the order authorizes mandatory government licensing or permitting requirements for AI model development and distribution. This language was clearly designed to address industry concerns about government overreach and regulatory burden.
Beyond the voluntary review process, the executive order also directs the Department of Justice to prioritize enforcement of crimes involving AI-assisted hacking and unauthorized computer access. This provision aims to address growing concerns about malicious uses of AI technology for cyberattacks and other criminal activities.
This isn’t Trump’s first attempt at comprehensive AI policy. In December, he signed an earlier executive order focused on creating a unified national AI framework. That order called for developing “one rulebook” to govern AI development across the country, with the goal of preventing a patchwork of conflicting state-level regulations.
The evolution of this latest order shows how the tech industry’s influence continues to shape AI policy at the federal level. Major AI companies have consistently argued that heavy-handed regulation could stifle innovation and give foreign competitors an advantage in what many consider the most important technology race of the 21st century.
