AIIncentives for U.S.-China Conflict, Competition, and Cooperation Across Artificial General Intelligence’s Five Hard National Security Problems
The prospect of either the United States or the People’s Republic of China —or both—achieving artificial general intelligence (AGI) is likely to heighten tensions and could even increase the risk of competition spiraling into conflict. But the emergence of AGI could also create incentives for risk reduction and cooperation. We argue that both will not only be possible but essential.
In a relationship marked by strategic rivalry and mutual suspicion, the prospect of either the United States or the People’s Republic of China (PRC)—or both—achieving artificial general intelligence (AGI) is likely to heighten tensions and could even increase the risk of competition spiraling into conflict.[1] This is unsurprising, as AGI could reshape the global balance of power or yield “wonder weapons” capable of overwhelming intelligence systems, information ecosystems, and cyber defenses (Mitre and Predd, 2025).
Yet the emergence of AGI could also create incentives for risk reduction and cooperation. We argue that both will not only be possible but essential. The United States and China will both want to avoid miscalculation and misunderstandings that could lead to an unwanted war. Neither will be able to manage alone the risks of AGI misuse—whether from rogue actors developing novel weapons of mass destruction (WMDs), autonomous agents triggering crises, or cascading disruptions that exceed state capacity. But such progress on risk reduction or on cooperation will not emerge organically. It will require a deliberate and carefully calibrated diplomatic effort to make it viable. In this paper, we offer some ideas for where to start, building on existing yet underdeveloped platforms and avenues for dialogue.
The PRC’s choices will be a principal factor in shaping the risks and opportunities that U.S. policymakers will face on the path to AGI, and, likewise, Chinese officials will likely view the United States as the most consequential external actor shaping AGI outcomes. At first glance, it might seem that the intensifying friction between Washington and Beijing over security, economic, and technology issues, along with each side’s extremely negative views of the other’s intentions, will drive relations between them in areas related to AGI.
However, the potential for the two countries to cooperate to reduce AI-related risks is illustrated by a November 2024 leader-level agreement that humans, not AI, should make decisions about using nuclear weapons (U.S. Embassy and Consulates in China, 2024). That agreement, though modest in scope, could serve as a foundation for future discussions about potential areas of mutual restraint related to the development and employment of AGI or other advanced AI capabilities. And so, the reality is likely to be much more complex, with choices between conflict, competition, and cooperation possible across this spectrum.
