The debate over artificial intelligence safety and the geopolitical contest over AI are usually covered as two separate stories. In an op-ed published this week, American Thinker contributor Dimitris Eleas argues that framing is a mistake — and that any American discussion about pausing or restraining frontier development has to account for what the country’s principal strategic rival is willing to do.
Eleas, a New York City-based political scientist and independent researcher, opens by pointing to a recent Wall Street Journal editorial on AI, which he says captured a critical point: if what developers see inside their own labs is genuinely dangerous, then responsibility and alignment work are warranted on their own merits. He pairs that with public remarks from Nvidia CEO Jensen Huang, who criticized Anthropic’s proposal and has said the goal is for America to succeed, and with Mark Zuckerberg’s framing that the United States should accelerate innovation while slowing geopolitical rivals.
From those threads, Eleas builds what he calls the paradox at the center of the AI debate. Restraint may be justified when frontier systems pose real risks, he writes, but restraint by American labs cannot be evaluated in a vacuum. The country’s chief strategic opponent is part of the equation whether policymakers like it or not. The piece’s central claim is not that safety should be abandoned, but that the choice between AI leadership and AI safety is a false one. America, in his telling, needs both.
No Putting It Back in the Box
Eleas is blunt about the irreversibility of the moment. The technology, he writes, cannot be recalled or locked away now that it exists. He also draws a distinction between what American companies might choose to do and what the broader competitive environment will permit. His argument is that when bitter rivals converge on the need for restraint, responsible restraint can happen — but it cannot be designed as though China will follow suit.
To make that point, he recites the scale of the challenge: a population of 1.4 billion, rare earth mineral reserves of 44 million tons, exports of $3.77 trillion, and government debt at 94 percent of GDP. The list is meant to convey the sheer weight of resources the Chinese state can bring to bear. In his view, the Chinese “do everything they can to beat America,” and the Communist Party’s centralized political system lets it steer enormous economic resources onto what he calls this new battlefield.
Whatever one thinks of that model of directed capitalism, Eleas argues, its ambitions in AI cannot be wished away — and he notes the same is true in space, where China has also made major advances. He describes the margin of AI leadership as potentially just a few months, because innovation is copied quickly. On the question of whether the U.S. and China could take joint action on AI risks, his answer is skeptical to the point of dismissal: only a naive person, he writes, would place absolute trust in a “red dictatorship where the bureaucrats reign.”
A National Security Question, Not Just a Tech Race
The stakes, as Eleas frames them, extend well beyond economics. He calls it a disaster and an act of hubris if China wins the AI race, and says AI leadership is increasingly tied to national security, military capability, scientific research, and the ability to shape the standards governing a technology that affects the entire world.
He concedes he rarely agrees with President Trump fully, but says Trump is right in saying the country must not slow down. Winning the AI race, in Eleas’s framing, is a way to protect a way of life — not merely a matter of market share or technological prestige.
Eleas also reaches for historical analogy. The internet itself, he notes, grew out of U.S. government-backed research, which undercuts any assumption that the next technological transformation must come entirely from the private sector. He suggests that a serious American frontier-AI research effort housed within the national-security establishment could produce capabilities and discoveries whose importance cannot yet be imagined — a case for public capacity that runs alongside, rather than against, private-sector leadership.
The Names Behind the Race
Turning to the executives he refers to only by first names — Dario, Sam, Elon, Jensen, and Mark — Eleas writes that they must know what to avoid so that AI does not become a Nemesis created by humanity’s own intelligence. The goal, he argues, should not simply be to build the most powerful system. It should also be to understand what must never be created, and to build safeguards that keep increasingly capable systems under human control.
That combination — acceleration plus clearly defined limits — is the balance he wants the people running the race for superintelligence to strike. He suggests they would do well to move closer to the spirit of the public statements made by Trump, Huang, and Zuckerberg on the subject.
The piece closes on a note of national purpose. Eleas writes that there is depth in the sacrifices made to forge the United States, and that decisions being made now about AI may eventually be judged as part of that larger American story. A multipolar world, in his view, will be far safer if the U.S. remains the strongest player — which makes American leadership on all fronts, in his telling, a necessity rather than an option.
The op-ed does not resolve the tension it identifies; it argues the tension itself is unavoidable. Safety advocates who want slower frontier development and hawks who want maximum speed are, in Eleas’s account, arguing about a question that cannot be answered by one country alone. His conclusion is that the United States should pursue capability and control simultaneously — because the alternative, in his framing, is to accept the premise that a strategic rival will exercise restraint even if America does.
Source: www.americanthinker.com — https://www.americanthinker.com/blog/2026/09/china-is-why-america-must-win-the-ai-race/