Trump AI Force Czar Race and Two Definitions of Regulation
The Space Race Refuses to Fade
In the 1960s, Washington and Moscow counted rockets against each other, and the count organized everything from school curricula to federal budgets. That frame is back, now applied to artificial intelligence, with the United States and China in the two seats once occupied by the US and the former Soviet Union. It survives because it does something the technical debate cannot: it turns a question about capability into a question about world order, and it makes speed look like security.
Several US companies are currently considered the industry leaders in AI, and the policy question is whether that lead can be held. The Trump administration intends to keep the regulatory burden light so that this stays the case. The operative policy question is therefore not liability, transparency, or consumer harm; it is whether a lead measured in months can be held at all.
The stakes are not only commercial. In Washington, the concern is that Chinese dominance in artificial intelligence could translate into Beijing’s rise as the sole superpower. Once the contest is framed in those terms, any slowdown reads as a concession.
Dario Amodei, CEO of Anthropic, took the logic to its end in a widely discussed essay on the risks of unchecked AI development. He warned of the consequences of continued unchecked AI development and outlined what the competition between the US and China could produce. If the United States greatly restrains its AI capabilities in the belief that China will do the same, and China then defects, the resulting systems could be powerful enough for that defection to lead to Chinese geopolitical dominance, he wrote. [1]. Any agreement, in his formulation, must therefore either have ironclad verifiability — the ability to independently check that an agreement is being honoured — or be limited enough that defection would not be militarily existential.
Free Rein Meets the Kill Switch
The optimist’s case and the skeptic’s case collided in the first half of September 2025. Several developers and decision-makers from major AI companies called for slowing the development of the technology. Some of the warnings came from developers about their own products. [1]. The backdrop was American concern that AI models were becoming too powerful — and a government unwilling to act. [1].
Trump’s answer was an all-caps post on Truth Social, the social media platform he owns. He called the idea that AI is taking over the world, destroying humanity, and all other things bad a “HOAX”. [1] He also asked when, in the history of business, anyone had seen the leaders of an industry call for regulation that, if strongly implemented, would drive them into oblivion and bankruptcy. [1]. In that reading, a warning from a company’s own engineers is evidence that the engineers have lost their nerve. [1].
Then came the reversal. Over the weekend, after yet another tirade about various alleged Democratic scams, Trump announced he would establish an “AI Force” and appoint a special envoy (a diplomat appointed for a specific, usually temporary, mission) to address the issue. He promised that he would by no means slow the industry’s growth. [1]. He said the government would look for “BAD”, and that it can do so very easily with the already existing criminal and civil justice system. [1] He added that he would announce the AI “Czar” (a senior government official appointed to coordinate one policy area) in the near future, and that only high-I.Q. individuals need apply. What duties and powers the position would carry, and what resources would be allocated to it, he left open. [1].
The timing points to a domestic rival. The previous day, California’s Democratic Governor Gavin Newsom had announced plans to establish a commission to regulate Silicon Valley-based tech companies. [1]. Among other things, Newsom wants to require those companies to build an emergency shut-off switch into their programs, which would allow humans to stop AI agents that have spun out of control. [1]. Several states, including some led by Republicans, are pushing for regulation. [1]. Trump’s announcement could be an attempt to undercut Newsom’s initiative.
Loyalty as the Regulatory Test
China has been regulating in this field since before the current American argument began. [1]. As early as 2017, the State Council of China — the country’s top administrative body, roughly the cabinet — issued a development plan for a new generation of artificial intelligence. [1]. Enforcement runs primarily through the Chinese Cyberspace Administration, the country’s internet and data regulator, which also issues licenses. [1].
The purpose those rules serve differs from the American debate. Regulation in China ensures that AI models do not run counter to official political doctrine. [1]. The question Chinese regulation starts from is how an AI system aligns with state objectives. [1]. That is a clear distinction from the risk-based regulation envisaged by the European AI Act — rules whose obligations scale with how risky an application is judged to be — and now being called for in the United States.
Thursday’s state dinner at the White House, where Trump will host Chinese President Xi Jinping, puts the two logics at one table. [1]. AI is likely to be very high on the agenda. [1]. The United States has already proposed establishing a mechanism for exchanging information about serious risks posed by AI. [1]. That proposal was discussed on Sunday in New York between Chinese Vice Premier He Lifeng and US Treasury Secretary Scott Bessent. [1]. Bessent said: “We think that, just like with any cross-border activity, that moving from opaque to more transparency between the number one and the number two AI powers in the world is very important.” [1]
Trade sits on the same table. [1]. For years, restrictions on high-performance AI chips and rare earth elements have been a recurring issue between the two countries, and the extension of a basic trade agreement is also on the agenda. [1]. Whether mutual trust is strong enough to lead to reciprocal concessions on AI will likely only become clear at the meeting.
Sources
1. Dw — Quote source (original article)
Mentioned organisations (context, not sources)
- Anthropic — Organisation (homepage)
- Truth Social — Organisation (homepage)
- State Council of China — Organisation (homepage)
- Chinese Cyberspace Administration — Organisation (homepage)
- European Union — Organisation (homepage)