President Donald Trump on Tuesday signed a revised executive order on artificial intelligence oversight, opting for a lighter-touch regulatory approach after significant pushback from the technology industry and prominent venture capital figures.
The signed order asks AI companies to voluntarily submit new models to the federal government for testing or evaluation 30 days before public release. This is a substantial retreat from an earlier draft that would have imposed mandatory pre-release review requirements on the most advanced AI systems, effectively giving the government a veto window over commercial AI launches.
The rollback came after organized opposition from Silicon Valley, including former White House AI czar and venture capitalist David Sacks, who argued that mandatory pre-release requirements would create a de facto licensing regime that could stifle innovation and disadvantage American AI companies relative to international competitors, particularly in China.
The final executive order includes explicit language clarifying its intent, stating that nothing in the order "shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement for the development, publication, release, or distribution" of AI systems. This language was inserted specifically to address industry concerns that a voluntary framework could evolve into a mandatory one through regulatory interpretation.
In a notable change of plans, Trump had originally scheduled a signing ceremony with top Silicon Valley executives in attendance, but ultimately signed the revised version privately after the more demanding draft was pulled.
Alongside the voluntary review provisions, the EO directs the Department of Justice to treat AI-assisted hacking and unauthorized access as high-priority enforcement targets, signaling that while the administration is backing off pre-deployment oversight, it intends to respond aggressively to AI-enabled cybercrime after the fact.
Why It Matters
The retreat to voluntary AI oversight reflects the current political reality in Washington: the technology industry retains significant leverage over AI policy, and mandatory pre-deployment review of AI models is off the table for now in the United States. This creates a notably different regulatory environment than the EU AI Act framework and may accelerate competitive pressure on European AI firms. For enterprises and AI developers, the voluntary-only regime means deployment timelines will remain fast and relatively unrestricted under the current administration.