The latest clash between Washington and Beijing over artificial intelligence reveals something bigger than a disagreement over AI safety. It shows that the race to dominate AI is rapidly becoming a new front in the strategic rivalry between the United States and China.

Anthropic CEO Dario Amodei’s call to “slow the pace” of AI development was presented as a warning about the dangers of increasingly powerful models. But Beijing clearly sees another dimension to the argument: any attempt by the United States to slow AI development while restricting China’s access to advanced chips and technology could also be an attempt to preserve America’s technological lead.

That is why China has rejected the warnings as “fearmongering” and described them as part of a Cold War-style strategy of containment.

There is a certain logic to Beijing’s argument. It is difficult to separate AI safety from geopolitics when the most powerful AI systems depend on advanced semiconductors, enormous computing infrastructure and access to cutting-edge technology. If Washington imposes tighter controls on China while simultaneously asking the industry to slow down, Beijing is unlikely to interpret that as a purely humanitarian effort.

For the United States, however, the argument is equally straightforward: allowing China unrestricted access to the technologies needed to build frontier AI could eventually create a strategic disadvantage with consequences extending far beyond the technology industry.

AI is no longer simply another commercial sector. It has potential applications in cybersecurity, intelligence, autonomous weapons, surveillance, scientific research, economic productivity and military decision-making. Whoever develops the most capable systems — and can deploy them at scale — could gain an advantage across multiple areas of national power.

That makes the AI competition fundamentally different from an ordinary technology race.

The danger is that the pursuit of strategic advantage could begin to undermine the very safety concerns that both countries claim to take seriously.

If American companies believe they must accelerate because China is accelerating, while Chinese companies believe they must accelerate because Washington is trying to contain them, both sides can become trapped in a technological security dilemma.

Each side sees its own actions as defensive and the other side’s actions as threatening.

That is precisely what made the Cold War so dangerous.

The irony is that the AI race may require more international cooperation at exactly the moment when geopolitical competition is making cooperation harder.

Amodei’s warning therefore presents Washington with a difficult choice. The United States wants to maintain its technological lead, prevent sensitive AI capabilities from strengthening a strategic rival and reduce the risks created by increasingly powerful systems. Those objectives do not always point in the same direction.

A complete slowdown could surrender America’s advantage. An unrestricted race could accelerate the development of systems whose risks neither government fully understands.

China faces a similar dilemma. Beijing has powerful incentives to close the technological gap with the United States, particularly as Washington tightens restrictions on advanced semiconductors and AI-related technology. But racing ahead without adequate safeguards could create risks for China as well as the rest of the world.

This is why the emerging U.S.-China AI rivalry should not be viewed simply as a contest over who builds the best chatbot.

It is becoming a contest over who sets the technological rules of the 21st century.

The United States wants to preserve its leadership and shape global AI standards around its own security and regulatory priorities. China wants to prevent those standards from becoming instruments of American technological dominance and to ensure that it has a major voice in determining how AI is governed globally.

The result could be an AI version of the arms race — except the weapons are algorithms, computing power, data and semiconductors.

And unlike the nuclear arms race, AI is being developed primarily by private companies as well as governments. That makes the situation even more complicated. Companies such as Anthropic and OpenAI may warn about safety risks, but they are also competing for enormous economic and technological advantages. Governments, meanwhile, increasingly view those same companies and their technologies as strategic national assets.

The central question is therefore no longer simply whether AI should be slowed down.

It is whether Washington and Beijing can find a way to compete without turning AI development into a race in which neither side feels able to apply the brakes.

For the United States, winning the AI race may mean staying ahead of China. But true leadership will require something more than building the most powerful models first. It will require demonstrating that technological superiority can coexist with responsible development and international rules.

For China, catching up with America may be a national priority. But treating every AI-safety proposal as merely an American containment strategy risks dismissing legitimate concerns that have nothing to do with geopolitics.

The greatest danger is that both sides may be right — and still make the world less safe.

America is right to worry about losing its technological edge. China is right to question whether restrictions presented as “security” also serve America’s geopolitical interests. But neither country can afford to assume that the other will eventually slow down.

That is the logic that turns technological competition into an arms race.

The future of AI may ultimately depend not on whether the United States or China wins, but on whether the two powers can prevent their rivalry from becoming a race toward capabilities that advance faster than their ability to control them.

Ramani Parkunan is a former journalist and media consultant who frequently covers digital transformation, cybersecurity, artificial intelligence, and corporate governance.