There’s A ‘Deeply Dangerous’ Arms Race Emerging On Global Stage

The United States and China are locked in an escalating artificial intelligence arms race, experts warn, as Chinese AI systems rapidly catch up to their American counterparts in a high-stakes bid for technological dominance.

Fresh metrics from Stanford’s AI Index Report 2025 show the U.S. lead in AI has all but vanished. Chinese AI systems now rival American models on key benchmarks and lead the world in research output, fueling anxiety in both Washington and Beijing — and sparking urgent calls to keep the rivalry from spiraling out of control.

“This arms race mentality is deeply dangerous,” Max Tegmark, an MIT physicist and prominent AI researcher, told the Daily Caller News Foundation. He described both nations as “barreling forward” out of fear of losing and edge. “If they keep cutting corners to outrun each other, we could all lose in the end.”

Tegmark notes that just a year ago, American AI systems were decisively ahead in key domains — for example, excelling at various performance benchmarks testing general knowledge and problem-solving ability — but that advantage evaporated in a matter of months, according to Stanford’s report.

“We’re basically neck and neck now,” Tegmark said, pointing out that Chinese models have matched U.S. models on major academic and coding challenges. “By some measures there’s virtually no difference anymore. That happened far faster than people expected, and it has major implications.”

Those implications extend to perceptions of national power. Advanced AI is widely seen as a cornerstone of future economic and military strength, and the prospect of parity — or Beijing overtaking Washington — has injected a new sense of urgency on both sides of the Pacific. Then-Sen. Marco Rubio, now secretary of state, framed the issue in terms of national security in a 2023 statement.

“The heads of Intel, Nvidia and Qualcomm are putting profits ahead of national security,” Rubio said. “These CEOs are lobbying against export controls designed to keep the Chinese Community Party away from AI technology. China is growing stronger by the day, aided by American capital and chips. We cannot allow China to get its hands on advanced AI chips.”

While U.S. tech firms still attract more private investment by a vast margin — some $109 billion in 2024, nearly 12 times China $9.3 billion, the Stanford report said — China’s AI labs have achieved world-class performance with far leaner budgets.

“It’s impressive and a little scary that China did so much with so much less funding,” Tegmark said. “It shows that talent and strategy are everywhere — a head start is never safe in tech.”

Indeed, China now leads in AI research papers and patents, Stanford reported, and its top companies — from Alibaba to Baidu — have rolled out AI models rivaling those from Google and OpenAI. The narrowing skills gap means the AI race is no longer about whether China can catch up, but how soon — and what it will do with a leading position.

“Whoever leads in AI will effectively write the global rules,” Tegmark observed. “Neither side wants to slow down, because they’re afraid the other side will get ahead — but speeding up without coordination is making the world less safe.”

Strategists worry the U.S.-China AI contest could reshape global power dynamics. Cutting-edge AI can confer advantages in everything from finance to cyberwarfare to autonomous weapons. American tech titans, notably Mosaic founder Marc Andreessen, likened the launch of China’s lean and effective DeepSeek R1 model to a new “Sputnik moment,” while China’s leadership has openly declared its aim to win the AI race by the end of the decade. Already, both governments are funneling resources into AI research and talent, driving a boom in domestic AI initiatives.

Deepseek R1 is AI’s Sputnik moment.

— Marc Andreessen (@pmarca) January 26, 2025

“By 2030, our country will reach a new world leading level in artificial intelligence theory, technology and application and become a principal world leader for artificial intelligence innovation,” Beijing declared in a 2017 statement, according to the Associated Press.

But Tegmark said the competition comes with risks.

“This isn’t just about who launches the best chatbot,” he said. ” It’s about who sets the agenda for how AI is used in the world — and that has military and geopolitical consequences.”

Amid the sprint for AI supremacy, global governance efforts are struggling to keep pace. Unlike the Cold War nuclear standoff — which eventually birthed treaties and U.S.-Russian communication lines — there is no binding international framework for AI.

Dan Hendrycks, head of safety at the Elon Musk-owned xAI, echoed the need for bilateral cooperation in a previous interview with the DCNF.

“We’re sleepwalking into a multipolar world with zero coordination,” Hendrycks warned. “If China, the U.S. and other can’t even talk about aligning incentives, it’s not a rivalry — it’s a shared crisis waiting to happen … Calling it an arms race makes it sound zero-sum. But AI safety isn’t like building nukes. It’s more like building aviation standards. Either the whole system works, or it doesn’t.”

In 2024, the United Nations began discussing principles for responsible AI use, and officials from both the U.S. and China have voiced support in abstract terms for AI safety. But tangible agreements remain elusive.

“We managed to avoid nuking ourselves by talking to each other and setting some ground rules,” Tegmark continued. “We need that same mindset with AI. Right now it’s a Wild West. Competition has to be balanced with collaboration and safety, or everyone could end up worse off.

Ethical concerns are also mounting as the rivalry accelerates. Both nations are deploying aI in ways that raise questions, from facial recognition surveillance to deepfake propaganda.

If the race continues unchecked, Tegmark and others fear a scenario where safety protocols are an afterthought. An AI system failing or behaving unpredictably — say, an autonomous weapon misidentifying a target — could spark crises. Yet each superpower worries that stringent constraints might hand the other side an advantage.

“It’s a classic prisoner’s dilemma,” Tegmark said. “Neither one of us wants to be the first to hit the brakes. The result is this turbo-charged development on both sides with almost no guardrails.”

The MIT physicist said the solution must ultimately be cooperative: unilateral restrictions are politically unpalatable, but a mutual agreement could ease the fear of falling behind. There is precedent for rivals finding common cause on existential risk, he notes, pointing to Cold War arms control.

For now, the AI race shows no sign of slowing — if anything, it is entering a new, more volatile phase as China closes the last gaps in capability. Tegmark’s prognosis is guarded.

“For the first time, humanity is racing to create something that could eventually outsmart us,” he said. “Whether this ends in a global AI renaissance or in disaster depends on choices we make now. I’d much rather see cooperation — even just basic communication — than a reckless contest to see who can deploy an uncontrollable AI first.”

Each breakthrough in machine intelligence, he argues, should heighten our collective responsibility to manage it safely. Instead, the U.S.-China rivalry has so far had the opposite effect: pushing both players to move faster. The world is watching closely, and anxiously, as the two AI superpowers charge ahead — “barreling toward a future,” Tegmark warned, “that we are not remotely prepared for.”

(Featured Image Media Credit: State Department Photo/Public Domain)