Beyond Killer Robots: How China Is Quietly Winning the AI War Without Firing a Shot

.

A Different Kind of AI Race

Most people picture an AI arms race involving killer robots or futuristic weapons. According to national security analysts, the real fight is already happening — and it looks nothing like science fiction.

Experts who study China's technology strategy say Beijing is running a quiet, three-part campaign against American AI leadership. It involves copying the "thinking" of top U.S. AI systems, undercutting American companies with free software, and eventually turning those tools toward political influence campaigns.

The warnings come at a pointed moment. Chinese President Xi Jinping is set to make his first U.S. state visit in more than a decade, arriving in Washington on September 24 for meetings with President Trump that are expected to include artificial intelligence as a central topic.

.

The Three-Part Playbook

Leah Siskind, an AI research fellow at the Foundation for Defense of Democracies, describes the situation less as a race and more as a drawn-out contest fought on multiple fronts at once — one where the U.S. is ahead in some areas and falling behind in others.

The first front is cost. American companies have poured enormous sums into building expensive, closely guarded AI systems. China, by contrast, has pushed hard to give away free, open-source alternatives worldwide. Because many companies simply cannot resist a free option, some are quietly shifting their operations onto Chinese-built models — without any real way to know whether those systems are secretly filtered by Communist Party censorship or contain hidden backdoors.

China analyst Gordon Chang argues Beijing is repeating a strategy it already used to dominate the global solar panel industry: undercut competitors on price until they are pushed out of the market, then take control once the field is clear.

.

Copying the Competition

The second front involves a technical shortcut called "knowledge distillation" (a method where a smaller AI system studies a larger one's responses to copy its abilities). Instead of spending billions to train an AI system from scratch, Chinese labs can extract much of that knowledge simply by extensively questioning U.S.-built models.

This is not a fringe theory. On September 8, 2026, the National Security Agency, the FBI, and the Cybersecurity and Infrastructure Security Agency (CISA) jointly confirmed the practice in an official advisory. The agencies named six Chinese firms — DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI — accusing them of running industrial-scale extraction campaigns against American systems, including models built by Anthropic, OpenAI, Google, and xAI, generating billions of exchanges since at least late 2024.

CISA's acting director called on American AI companies to urgently defend their platforms, warning that the practice is closing the gap U.S. firms worked for years to build. The 2025 release of China's DeepSeek model, which stunned the industry by matching top American systems at a fraction of the training cost, is widely seen as an early result of this strategy.

.

Beyond Software: Robots, Factories, and Propaganda

The threat isn't confined to computer code. Siskind points to what she calls "embodied AI" — artificial intelligence built directly into physical infrastructure such as factory robots, commercial drones, and automated supply chains — as an area where China is moving fast and largely out of public view.

Chang raises a further concern: AI-enhanced propaganda designed to shape public opinion and influence elections. He argues that a China empowered by advanced AI could pursue outsized global influence, comparing it to the ancient Chinese imperial concept of "Tianxia," or the idea that a ruling power is entitled to guide the world.

.

America's Advantages — and What Washington Should Do

The United States still holds real advantages. The country operates more than 5,400 data centers compared to roughly 450 in China, alongside a significant lead in advanced computer chips.

To protect that edge, Siskind argues Washington needs to close loopholes in chip export controls, blacklist foreign labs caught running deceptive extraction campaigns, and invest far more heavily in American open-source AI development — so that the most capable AI systems are built under democratic, not authoritarian, oversight.

.

Trump Doubles Down as Xi's Visit Approaches

President Trump has made clear he intends to keep America in the lead. Addressing the United Nations General Assembly this week, he announced that U.S. government documents will now refer to advanced AI as "superintelligence," declaring that America currently leads China and the rest of the world "by a lot" and vowing to keep it that way. He also rejected any global effort to place artificial intelligence under centralized international control.

In a separate segment on Fox News' "America Reports," correspondent Steve Harrigan underscored the stakes of the moment, framing the U.S.-China AI competition as a race where the ultimate winner stands to gain outsized global leverage — a point Trump himself has echoed repeatedly in recent public remarks.

With Xi's state visit now just days away, artificial intelligence is expected to sit near the top of the agenda between the two leaders — a rare case where a quiet, largely invisible technology race could shape the outcome of very visible, high-level diplomacy.

.

Sources

1. National Security Agency, CISA, and FBI — Joint Cybersecurity Advisory AA26-251A (via Nextgov summary): https://www.nextgov.com/artificial-intelligence/2026/09/intelligence-agencies-warn-chinas-large-scale-ai-model-distillation-efforts/415851/
2. Reuters — "Trump says Xi will visit US on September 24": https://www.internazionale.it/ultime-notizie-reuters/2026/07/23/trump-says-xi-will-visit-us-on-september-24
3. Unite.AI — "NSA, CISA, FBI Warn China-Based AI Firms Distill US Frontier Models": https://www.unite.ai/?p=463550
4. Fox News — "President Trump warns whoever wins AI race 'has it all'" (video segment, background reference): https://www.foxnews.com/video/6405453669112

.