The Trump administration's initiative to establish a new regulatory body for artificial intelligence through an executive order has reportedly stalled, with the proposal facing significant internal friction. According to sources familiar with the matter, a draft executive order has been circulating within the administration in recent weeks, suggesting the creation of an industry self-regulatory organization for companies developing cutting-edge AI models. While the drafting of this order represented a key step toward establishing such a watchdog, major obstacles remain.

Three additional insiders indicate that finalizing the order would require backing from several senior officials, including the president, yet momentum has recently ground to a halt. The concept for this new regulator stems from a proposal made last month by Demis Hassabis, the former CEO of Google DeepMind, who advocated for a novel public-private entity to oversee AI. Hassabis took to X to call for a standards-setting body modeled after the Financial Industry Regulatory Authority (FINRA), a self-regulatory organization that has long governed U.S. brokerages and securities professionals.

The push for this executive order also draws from various suggestions put forward by senior administration figures, with reports from last month indicating that Treasury Secretary Scott Bessent had pitched the creation of such a regulator. The reasons behind the order's current stagnation remain unclear, though the FINRA-style approach for an AI body has drawn sharp criticism from David Sacks, the former White House AI and crypto czar. On his "All-In" podcast last week, Sacks dismissed the idea as a "terrible idea," taking particular issue with pre-release model testing and the notion of a self-regulatory body reporting to the government. He instead voiced support for a model akin to the Motion Picture Association, which independently runs the film rating system without government oversight.

Further specifics on the executive order or the proposed self-regulatory organization are not yet available, and its timeline remains uncertain. In June, the administration issued a separate AI-related executive order that established a voluntary framework for evaluating frontier models, a document that took months of negotiation to finalize. A White House spokesperson remarked, "Any policy announcements will be made directly by the president. Discussions about potential executive orders are purely speculative."