Image Courtesy: Getty Images

A former Obama administration national security official has argued that the United States should prepare for the possibility of military action against Chinese artificial intelligence infrastructure if Beijing gains a decisive lead in developing artificial general intelligence, or AGI.

Jacob Stokes, now deputy director of the Indo-Pacific Security Program at the Center for a New American Security, outlines the proposal in a report titled “Superpowers and AGI.” The report examines how the United States might respond if China were to achieve AGI first, including through cyberattacks, sabotage and, in the most extreme scenario, physical attacks on data centers.

AGI is a theoretical form of artificial intelligence capable of matching or surpassing humans across a broad range of intellectual tasks. Although there is no established definition or consensus on whether AGI can be achieved, Stokes’ analysis treats its eventual development as a possibility that national security planners should prepare for.

In one scenario, Stokes suggests that a United States falling behind China could attempt to disrupt its rival’s AGI systems through physical infiltration or offensive cyber operations. He notes that cyberattacks could potentially remain below the threshold of conventional warfare, although their consequences would be difficult to predict.

A more serious option would involve direct attacks on the physical infrastructure supporting advanced AI systems. Stokes describes kinetic attacks, including the use of missiles and bombs against data centers, as the most dangerous option because of the risk of triggering a major escalation between Washington and Beijing.

“Data centers can be bombed with conventional munitions,” Stokes writes, while acknowledging that such an attack would cross a significant threshold in U.S.-China relations.

He argues that policymakers should examine these possibilities in advance so they can better understand the strategic implications of a technology that could become as consequential to national security as nuclear weapons were during the Cold War.

The proposal comes amid intensifying competition between the United States and China over AI development and computing infrastructure. Chinese companies have made rapid advances in AI, while U.S. firms continue to lead many major language-model benchmarks and have access to substantial computing resources.

Stokes warns that a sudden technological breakthrough by China could accelerate its development of advanced AI capabilities, potentially giving Beijing access to sophisticated cyber capabilities and other military advantages.

However, the prospect of bombing data centers remains highly theoretical. No evidence currently establishes that China has achieved AGI, or that destroying AI infrastructure would prevent such a system from being developed elsewhere. Any military strike against China would also carry consequences far beyond the AI sector, including the risk of a broader armed conflict between two nuclear-armed powers.

The debate highlights how rapidly AI has moved from a technological competition to a national security concern, even as the capabilities at the center of that competition remain uncertain.