On September 19, 2026, Donald Trump rejected warnings about artificial-intelligence risks as a “hoax” while proposing an AI Force modeled on the Space Force and promising a future AI czar. The announcement put a new government structure beside an intentionally pro-growth message: the United States should keep advancing AI rather than slow down and risk losing ground to China.
The AI Force is a proposal, not an operating federal institution. Trump described the idea and its broad purpose, but gave no public operating structure in the announcement. He also said an AI czar would be announced in the near future without naming an appointee.
Trump proposes an AI Force and future AI czar
Trump presented the proposed force as a way to identify harmful AI activity while helping the industry expand. His comparison with the Space Force suggests a dedicated government organization, but the announcement did not turn that concept into a functioning agency or military branch.
The accompanying promise of an AI czar would give the administration a single senior official focused on the technology. For now, the role remains a planned appointment rather than a named postholder.
Trump’s position combines encouragement with a law-enforcement backstop. On September 21, he said the Department of Justice and other law-enforcement bodies could rein in harmful AI activity if necessary. That approach relies on existing criminal, civil and regulatory powers instead of describing a new, dedicated AI statute.
What Trump says the force would do
Trump said the proposed AI Force would look for “bad” activity and help watch over the industry as it grows. His message was not a call to freeze research or restrict the sector broadly. It was a promise to support expansion while intervening against conduct the administration considers harmful.
He also predicted that AI could eventually account for 25% of U.S. gross domestic product. That figure is Trump’s prediction, not an established economic forecast.
The policy logic is straightforward: rapid development is treated as a strategic asset, while criminal and civil enforcement would address abuses after they emerge. The unresolved question is how that system would identify risks early enough to act before they become operational or social damage.
Why the announcement arrives during a safety dispute
The proposal landed amid a disagreement inside the AI industry over whether frontier-model development is moving faster than safety work. Dario Amodei, Anthropic’s CEO, has argued for a slower pace so monitoring and safeguards can keep up. Sam Altman has been reported as supporting additional caution and oversight at OpenAI, while other technology leaders have also backed stronger controls.
The divide is not simply “AI or no AI.” It is about the speed of capability growth, who evaluates the systems and whether companies should follow shared rules even when competitors keep moving.
Daniel Kokotajlo, a former OpenAI researcher, offered one of the starkest counterarguments to Trump’s dismissal. He estimated a 70% chance of a globally catastrophic AI outcome. That number is Kokotajlo’s personal estimate, not a consensus scientific measurement.
The interview with Kokotajlo also presents his view that AI research could become increasingly autonomous and that superintelligent AI could arrive by the end of the decade if development continues at its current pace. Those are forecasts about a possible future, not established outcomes.
The U.S.-China competition behind the argument
Trump has framed calls for slower development as a potential strategic gift to China. In that view, the United States needs to preserve its lead in AI, and excessive restrictions could weaken its economic and technological position.
That argument turns safety policy into a competition policy. A safeguard is no longer judged only by whether it reduces risk; it is also judged by whether it gives another country an advantage. Trump’s proposed AI Force fits that framework by pairing oversight with an explicit commitment to industry growth.
The result is a tension that will shape any future policy: Washington could seek stronger monitoring while still trying to accelerate American AI development. A government office focused on harmful activity would need to operate within that balance rather than simply choosing between unrestricted growth and a halt.
The risks beyond superintelligence
The safety debate covers more immediate problems than a hypothetical loss of control by superintelligent systems. Reported concerns include algorithmic bias, privacy violations, misinformation and failures of accountability. They also include job displacement, rising energy demand, water use, pressure on electricity grids and local costs associated with data centers.
These risks operate on different timescales. A catastrophic scenario is a forecast about a possible future. Privacy failures, misleading automated content, workplace disruption and infrastructure demands are present policy questions that can affect people and communities without requiring superintelligence.
That distinction explains why some technology leaders support additional safeguards even while continuing to develop AI. Their position is that progress can continue, but evaluation, monitoring and accountability must grow alongside capability.
The proposed AI Force would therefore face two different jobs: respond to harmful activity under existing legal powers and help define what deserves attention before the harm occurs. Trump’s September 19 announcement supplied the political direction; the practical design of that response remains the next step.
