In the US-China AI race, the real fight is keeping humans in control

The US-China AI competition faces a dilemma. Both countries see artificial intelligence as a source of economic, scientific and strategic power, so neither has much incentive to slow down. Yet the more capable and autonomous AI becomes, the more important the question of human control.

Last week’s summit between US President Donald Trump and Chinese President Xi Jinping brought that tension into the open. Trump wants to leave AI “exactly where it is” rather than add new guardrails, while Xi believes China and the United States have a responsibility to develop and manage AI for good and ensure that it remains under human control.

This is more than a disagreement over regulation. It reveals a deeper dilemma facing both powers. Competition rewards faster deployment and greater delegation to machines. Control requires limits, monitoring and the ability to intervene. The same systems that increase national power can therefore make human authority harder to preserve as AI becomes more autonomous.

Conventional geopolitics treats AI mainly as an instrument: the US builds better systems, China builds better systems and each side uses them to advance human-defined objectives. But what happens when the instrument can act for sustained periods without continuous human direction? AI could then become an independent agent in international politics, not merely a tool used by states.