Trump's 'AI Force' and AI Czar Put Washington Squarely on Team Acceleration

The White House has picked its side in the AI safety fight, and it picked in public. On Saturday, President Donald Trump announced on Truth Social that he is forming an "AI Force" — explicitly modeled on Space Force — and will name an artificial intelligence "czar" in the near future, while promising the industry the one thing safety advocates spent the summer asking Washington to consider imposing: nothing at all. "We will not in any way hinder or stifle the Growth of this incredible Industry," Trump wrote (Fox News). The announcement answers the noisiest AI-risk debate of the year with a slogan, a vacancy, and a promise that no lab will be slowed down. For builders, the operative detail is what Washington just assigned to you instead.

What Actually Happened

The announcement came in a lengthy Truth Social post on Saturday, September 19. Trump framed the government's role as protection plus watchfulness: the administration would resist efforts to restrict the technology while using "the existing criminal and civil justice systems" to address wrongdoing, adding that the government would be "looking for BAD" (Fox News). "For this purpose, I am forming the AI Force, much like I did Space Force, which has been a tremendous SUCCESS, in my First Term," he wrote, followed by the czar tease: "I will be announcing, in the near future, the AI 'Czar' — Only High I.Q. individuals need apply!" (Fox News).

What the post did not include is the story. No structure, no authority, no budget, no membership, no candidate, and no timeline beyond "near future." Fox News reported that Trump "did not provide details about the structure, authority or membership of the proposed AI Force or identify potential candidates." The Washington Post's account led with the same absence: an unspecified "AI Force" and a czar announcement to come, paired with an explicit rejection of calls from lawmakers and industry leaders for new constraints (Fox News, WaPo).

The czar seat itself has been empty since March, when David Sacks left the White House AI and crypto post; Sacks still advises Trump as co-chair of the President's Council of Advisors on Science and Technology (Yahoo). That vacancy is why the announcement matters more than its content: the administration currently has no dedicated AI point person, and the pick will define how "watch over it" translates into procurement, standards, and enforcement.

The post also carried the argument Trump has been making all month — that the AI backlash is manufactured. He filed "the decimation, or destruction, of AI" alongside his list of what he calls hoaxes, argued the effort started as an attack on data centers and fizzled once communities saw the economic upside, and claimed AI could eventually account for as much as 25% of US GDP while the US leads China in building it (Yahoo). Per Yahoo, he called fears of an AI takeover a hoax during a live call with Nvidia CEO Jensen Huang last week.

The timing was not accidental: the post lands amid calls from Anthropic's Dario Amodei, OpenAI's Sam Altman, and other industry leaders to slow frontier-model development, accelerating state-level rulemaking, and a weekend of congressional guardrail arguments (Fox News, WaPo). Trump's answer to all of it, in one post: full speed, existing laws, and a czar to be named.

Washington's Position Is Now Testable: Watch, Don't Slow

The administration's Sunday messaging made the policy concrete, and it is more coherent than the post's aesthetics suggest. Michael Kratsios, director of the White House Office of Science and Technology Policy, said on Fox News Sunday that companies worried about their own models should act on that belief unilaterally: "If you do believe that you're developing a technology that is unsafe, or you don't want it out into the world, you can just stop it. You don't need someone to force you to do that" (Fox News).

Kratsios also argued the federal government already has the tools it needs, citing an incident this summer in which, he said, the government received credible evidence of cybersecurity issues involving a model: "A letter was sent to them, and the model was pulled down until those cybersecurity issues were resolved." He rejected the runaway-AI framing outright, saying fears of "Terminator running on the streets" are the wrong message to send Americans. The broader posture predates the weekend: a June national security directive called for faster AI adoption while requiring systems used by the national security enterprise to be reliable, robust, steerable, and controllable (Fox News).

That is the regime in practice — existing authorities, targeted interventions after the fact, and no prospective rulemaking. Whether it is adequate is exactly what the other side of the debate disputes, and the dispute is no longer theoretical. Senator Peter Welch (D-VT), also on Fox News Sunday, made the structural argument: "Should banks regulate themselves? Should security companies regulate themselves? There's a public interest here in safety." Senator Dave McCormick (R-PA) took the industry-leads position but attached a condition with teeth: companies must bear liability for the risks they create. "You can't stand up in front of the public and the shareholders and say, on one hand, well, we're gonna raise hundreds of billions of dollars... and at the same time, say it poses a risk of destroying humanity. The responsibility is with the companies. And that liability is with the company" (Fox News).

The Industry Split Behind the Announcement

The acceleration camp now has its business case-maker on the record too. In a CBS News interview published Sunday, Nvidia CEO Jensen Huang put a number on the extinction debate: "2030 is not going to be the end of the world. There is 0% chance that's going to be the end of the world." Scaring people, he said, is "unnecessary" and "irresponsible," and predictions like them are "not grounded in science" (CBS News via Fox News, Tom's Hardware).

Huang was responding to recent claims from former Anthropic researcher Jacob Coxon, who wrote on social media that AI developers "earnestly believe that it could kill us all by the end of the decade" (Fox News). Yahoo reported that Marcus Williams of OpenAI's Safety Oversight team has put extinction odds at 70% within three years without regulation or a coordinated slowdown, and that Bilal Chughtai quit DeepMind's AGI safety team on similar grounds. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman are among the industry leaders who have called for slowing frontier-model development and for more federal regulation (Fox News).

Huang answered the conflict-of-interest question directly. Asked about Nvidia's financial interest in continued AI development — the company makes the chips every frontier lab trains on — he said: "Our company's success is directly connected to the safe deployment of products and services. If we don't continue to do that, our value would be diminished." And on regulation: apply existing law first. "Existing laws covering areas such as cybersecurity, unauthorized access and liability should be applied" before new AI-specific rules (Fox News, CBS News).

That is the same position as the White House's, arrived at from the other direction: existing law, company discretion, no new rules. McCormick's liability argument is the pressure point in both. If the federal government declines to write AI-specific rules, liability allocation becomes the mechanism that disciplines the industry — negotiated in contracts, litigated in courts, and priced by insurers rather than set by an agency.

The China Track the Announcement Didn't Mention

While Washington was promising the domestic industry it would not be slowed, Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng were scheduled to meet in Manhattan on Sunday — alongside US Trade Representative Jamieson Greer — with artificial intelligence on the table alongside tariffs and critical minerals, according to Reuters (Fox News). The AI discussions cover both open- and closed-weight models, as Washington and Beijing consider ways to address shared risks from increasingly powerful systems, ahead of a planned Trump-Xi summit.

Bessent's framing: "The United States remains the leader in AI. And we are open to discussions on avoiding shared risks and avoiding bifurcation of our two systems." He has previously called for US-China AI guard rails to keep powerful models out of the hands of malign non-state actors. The talks connect directly to hardware: rare earths are central to manufacturing the advanced semiconductors that power AI, and Chinese open-weight models are gaining traction among US companies (Fox News, Reuters).

The weekend thus produced two federal postures at once. Domestically: no constraints, no slowdown, a czar pending. Externally: negotiators discussing AI guard rails with the competing superpower. Both can be true — but builders selling into either ecosystem should note that the acceleration promise has an exception already carved out, and it's spelled "China."

What Builders Should Take From It

• The federal compliance regime is now, officially, "existing law plus your own judgment." Don't budget for a new federal AI approval process — budget for the liability, procurement, and incident-response conversations that existing cybersecurity and product-liability law already imply.
• Liability is being deliberately parked on companies. McCormick said the quiet part on Sunday, and Kratsios's "you can just stop it" confirms the philosophy. Your eval evidence, incident documentation, and model-change controls are now sales assets — buyers will ask for exactly what Washington declined to mandate.
• The czar pick is the signal to watch. The role has been vacant since March. Whoever fills it decides how "cherish it, help it, and watch over it" becomes procurement rules and enforcement priorities. If you sell AI into federal programs, that appointment is your roadmap.
• US-China guard-rail talks put model sourcing inside trade policy. If you build on Chinese open-weight models, note that bifurcation now has negotiators attached — watch what Bessent trades before the Trump-Xi summit.
• Price the doom debate by incentive. The chipmaker says the risk is 0%, the labs that build the models say it may be existential, and the White House says it's a hoax. None of these positions is disinterested; none of them is your risk model either.

Developer312 covers the AI business signals builders actually need to act on. Get the weekday briefing at developer312.com.