AI Arms Race: RAND Expert Warns How China Competition Is Reshaping US National Security
The New Battlefield
Artificial intelligence is no longer just a commercial gold rush—it has become the central front in the strategic competition between the United States and China, fundamentally altering calculations around national security. According to Matan Chorev, vice president and director of RAND Global and Emerging Risks, the speed of AI adoption is creating unprecedented vulnerabilities for the U.S., extending far beyond a simple race for the most advanced chatbot. In a new Q&A, Chorev lays out a landscape where asymmetric advantages, strategic surprise, and systemic risks are redefining military and intelligence planning.
Beyond the Model: Systemic Vulnerabilities
Chorev’s assessment frames AI as a multi-domain accelerator that intensifies competition across military, intelligence, cyber, and industrial sectors. While much public debate focuses on model capabilities—such as large language models powering disinformation—the deeper security risks lie in the infrastructure. Chorev points to fragile supply chains, data dependence, and the global talent competition as critical weak points. The risk is not merely that China might build a superior weaponized algorithm, but that the U.S. defense ecosystem could find itself dependent on adversarial-controlled hardware or software components embedded in sensitive systems.
The conversation highlights how AI is changing the nature of decision-making under pressure. Enhanced targeting systems, pervasive surveillance powered by computer vision, and autonomous cyber operations can compress the decision cycle, leaving human operators struggling to keep pace. This compression raises the specter of inadvertent escalation, where automated systems react faster than political leadership can intervene, a scenario that looms large in tensions over Taiwan or the South China Sea.
The Governance Tightrope
A core challenge identified in the Q&A is the delicate balance between innovation and security. Policymakers must encourage the rapid domestic development of AI to maintain a competitive edge while simultaneously limiting exposure to foreign dependence and weak safeguards. Chorev suggests that treating AI purely as a commercial technology race misses the point; it is fundamentally a contest for strategic power. This requires the U.S. to secure its innovation base without stifling it, ensuring that the next generation of defense systems does not carry hidden digital Trojan horses.
“The risks are not limited to model capabilities themselves, but also to supply chains, data dependence, talent competition, and the deployment of AI in sensitive systems,” Chorev explains, underscoring the need for a holistic view of the threat landscape. This includes evaluating how AI-fueled information warfare can undermine social cohesion during a crisis, a domain where China has demonstrated increasing sophistication.
Implications for Deterrence
RAND’s analysis ties these concerns directly to U.S. foreign policy and defense posture. China’s rapid AI development, driven by massive state investment and vast data pools, creates challenges for strategic stability. The integration of AI into conventional and nuclear command-and-control systems could alter deterrence calculations, making it harder to signal intent clearly. Chorev warns that avoiding strategic surprise requires the U.S. intelligence community to not only monitor adversary capabilities but to deeply understand how AI is being integrated into their command structures—a task made difficult by the opacity of closed systems.
The Q&A serves as a clarion call for a whole-of-government approach, linking broader AI governance questions with concrete defense planning. As the U.S. government grapples with export controls on advanced chips and the regulation of frontier models, the underlying message from RAND is clear: the window for setting the terms of AI security is narrowing, and the cost of complacency is a widening gap in strategic readiness.
