Trump Announces an AI Force and AI Czar, Dismissing Safety Fears as a Hoax
Trump pledged in a Truth Social post to create an AI Force and name an AI czar, dismissing safety fears as a hoax, days after every frontier lab agreed to slow development and accept embedded evaluators.
The political response to this month’s safety cascade arrived on September 19, and it is the opposite of what every frontier lab asked for. Per Bloomberg, Trump pledged in a Truth Social post to create an AI Force and name an AI czar, dismissing safety fears as a “hoax” and rejecting any slowdown. The announcement came within days of all four frontier lab chiefs publicly agreeing to pace development and accept embedded evaluators, making the White House position unmistakable: the US government will not be the mechanism for restraint.
What the AI Force Actually Means
The details are thin beyond the pledge, and the closest precedent is the Space Force model: a dedicated organizational unit rather than a regulatory body. The AI czar appointment fills the role the Biden administration left vacant after the AI Safety Institute was folded into NIST’s broader mission. What is clear from the framing is that the AI Force is an acceleration instrument, not an oversight one, which matches the White House’s June executive order disclaiming mandatory licensing and the G20’s light-touch principles the US championed.
The Timing Tells You What It Responds To
The announcement did not happen in a vacuum. This week alone: all four frontier CEOs agreed to pace development, Amodei proposed embedded evaluators and a global coordination ladder, the Pentagon’s AI-assisted intelligence nearly triggered a war with China, California signed first-in-the-nation AI safeguards, and the Sanders-Casar bill proposed criminal penalties for superintelligence development. The AI Force announcement is the counter-framing to all of it simultaneously: the federal government will organize for AI speed, not AI restraint.
What to Watch
The practical question is whether the AI Force absorbs or competes with the existing safety infrastructure (the AI Safety Institute, CISA’s AI guidance, the FRONTIER Act’s proposed Under Secretary). The antitrust question from OpenAI’s slowdown request also lands here: an administration that rejects safety fears as a hoax is unlikely to grant the waivers labs need to legally coordinate restraint. Meanwhile, the BRICS counter-bloc and the EU’s enforcement model are filling the governance vacuum from opposite ends. The AI Force is not the end of the pacing debate; it is the moment the debate became fully partisan.
Get Insanely Good at AI
The book for developers who want to understand how AI actually works. LLMs, prompt engineering, RAG, AI agents, and production systems.
Keep Reading
gr.Workflow Turns AI Pipelines Into Deployable APIs
Learn how to model, debug, expose, and deploy multi-step AI pipelines with Gradio Workflow and daggr.
The AI Pacing Plan Just Triggered a Three-Way Geopolitical Fight
Trump called AI guardrails a sick conspiracy and said only a High IQ president is needed, Beijing branded Amodei's pacing plan a Cold War tactic while pitching a BRICS counter-bloc, and AI stocks slumped on the slowdown calls.
Trump Dismisses AI Extinction Warnings, Pledging to Keep the China Lead
President Trump said he has no concerns about AI leading to human extinction and that the US leads China in AI by roughly a year, publicly rejecting the risk estimates that Anthropic and OpenAI staff endorsed this week.
G20 Unanimously Endorses the US Light-Touch Carolina Principles on AI
All G20 members, including China, endorsed the non-binding Carolina Principles for AI governance at the Chapel Hill ministerial on September 2, favoring sector-specific rules over comprehensive regulation.
Anthropic and Accenture Commit $2 Billion to Embedded AI Evaluators
Anthropic and Accenture announced a $2 billion five-year partnership embedding Accenture evaluators inside Anthropic with employee-level access throughout the model development lifecycle, the first commercialization of the embedded-evaluator model.