More than 60 allies of President Trump, led by former strategist Steve Bannon and conservative activists Amy Kremer and Brendan Steinhauser, are pressing the administration to test and approve advanced artificial intelligence systems before they reach the public. The letter, obtained first by Axios and organized by the group Humans First, marks a sharp break from White House policy.
Inside the administration, the dominant view favors minimal regulation, with officials arguing that light-touch rules will help the U.S. win the global AI race. Even those who support model evaluation have resisted the idea of mandatory government sign-offs, creating a deepening rift between the White House and a vocal faction of its base.
Bannon, who hosts the influential "War Room" podcast, has warned for over a year about potential job losses from AI. "This letter takes us next level," he told Axios, adding that it demands "mandatory testing and government approval" of powerful systems. The group Humans First, whose tagline says "technology should serve humans ... not replace them," organized the effort.
If adopted, the proposal could slow the release of cutting-edge AI models in the U.S., potentially ceding ground to foreign competitors. It also threatens to upend the industry's current development pace, where companies like OpenAI and Google often deploy updates with little pre-release oversight from Washington.
The letter's timing is notable as Trump weighs AI policy ahead of the next election cycle. Critics within the administration argue that government pre-approval would stifle innovation, while Bannon's faction contends it is necessary to protect American workers from looming disruption.