The AI Revolution and Global Order

by Hal Brands

Technological revolutions make geopolitical revolutions. The Industrial Revolution brought an era of European global supremacy over weaker, less-developed societies. The rise of mass manufacturing transformed the United States from an agrarian nation into a global superpower. The nuclear revolution crystallized a bipolar Cold War order. The digital revolution then helped the West exhaust and defeat the Soviet Union; it ushered in an era in which America had no peers.

Technological upheavals shake the foundations of economic and military power. They unsettle existing hierarchies and reset established relationships. They deliver creation and disruption on a massive scale. Now, the world is embarked on another technological transformation with radically disruptive potential—the revolution in AI.

The first months of 2026 brought striking illustrations—from a daring U.S. raid into Caracas to a devastating U.S.-Israeli assault on Iran—of how AI-enabled capabilities are making warfare faster, more lethal. Intelligence collection and analysis, cyber operations, and other tools of statecraft are being reshaped in real time.

Ahead, perhaps, lies a breakthrough to AGI, or artificial general intelligence—machines capable of doing complex tasks as well as, or better than, humans across a broad range of activities. That breakthrough, if it happens, might unlock even more dramatic economic and military gains for its winners. Or it might bring catastrophic instability to societies around the globe. One way or another, the interaction between AI and geopolitics will be a defining theme of our age.

Will AI be a force for freedom or a tool of tyranny? Will it concentrate or diffuse global power? Will it foster greater international stability or sow devastating upheaval? How will it change the conduct of war and diplomacy? Will AI ultimately present such existential challenges that it compels states to look past old patterns of rivalry? Perhaps most important, what strategy should America pursue in the age of AI?

These are among the questions RAND’s Center for the Geopolitics of AGI has wrestled with since its inception. They were also the subjects of a conference I hosted at Johns Hopkins University in April, which produced a collection of essays—titled Geopolitics of AI—being published by Johns Hopkins University Press later this month. As the name implies, that book owes a tremendous debt to ideas and individuals associated with RAND’s work. Indeed, several contributors to the volume have been involved with RAND’s Geopolitics of AGI project in one way or another.

The essays in the forthcoming volume offer a wide range of perspectives and conclusions. In this post, I preview five interesting contributions and themes.

The first theme, offered by Salman Ahmed and Matan Chorev, is that AI may bring a new bipolarity—but even the world’s strongest actors will have to navigate their dependence on others. The reason has to do with the structure of the global AI stack—the location of resources and inputs that enable AI innovation. Only America and China have the full suite of capabilities necessary to compete at the frontier. One possible future could see countries everywhere become takers of tech made by the two greatest powers: There won’t be true “strategic autonomy” for lesser powers. Yet even Beijing and Washington will need access to tools, resources, and expertise provided by weaker states, from the UAE to Singapore. Success is thus a matter not of achieving full technological independence, but of building the strengths and coalitions that allow one to favorably shape the inevitable dependencies.

A second theme, advanced by Michael Sulmeyer, is that AI could provide a powerful boost to cyber offense against cyber defense—albeit for reasons that are more institutional than technological. Strong AI could enable more effective performance at every stage of the AI kill chain—the steps required to prepare and execute offensive operations. There’s no technological reason why AI can’t also empower cyber defense. But the policies and practices necessary for that to happen—from the automation of threat removal to the integration of America’s institutionally fragmented cyber defenses—are lagging badly behind the threat. Unless the United States adjusts, rapidly, to this challenge, it will be at risk of exploitation by bad actors ranging from cyber criminals to rival states.

A third theme, emphasized by Michael Mazarr, is that AI may accelerate and intensify strategic interaction. Mazarr models a scenario in which rival AI models are responsible for managing a U.S.-China confrontation in the South China Sea. His chapter previews a future “world of unrelenting, high-speed, high-intensity strategymaking in which AI agents trade dozens of moves a day, probing and experimenting to find advantage.” Policymakers may understand that such rapid, back-and-forth interactions bring a heightened risk of escalation—but they may worry that not automating key decisions risks leaving them a poor second against faster rivals. The ultimate risk, then, is that the rise of machine-speed strategy may bring a loss of human control in perilous standoffs.

A fourth theme, which runs through several chapters, is that success in the AI era requires far more than just having the best tech. The United States could have the most advanced frontier models but still lose the strategic contest for influence if China’s cheaper open-weight models conquer the world. Moreover, countries that thrive in the AI era will be those that invest in the deeper, more fundamental sources of national competitiveness: the capacity of their institutions to absorb and diffuse new technologies, the resilience of political systems that must encourage innovation while weathering technological disruption, the ability to preserve a healthy, balanced relationship between the private sector and the state, the health of the international coalitions that serve as force-multipliers for national strength. AI may be a revolutionary technology, but so many core ingredients of national power remain the same.

The global impact of AI is just starting to unfold, of course, which relates to the final theme—the importance of getting the big questions right. The age of AI is, for now, the age of uncertainty: The future evolution of the technology and many of its geopolitical implications remain hazy. That technology will confront policymakers with hard questions about war and peace, power and influence, the relationship between humans and machines. It is crucial, in these circumstances, that there are intense intellectual debates about the contours of the coming age and the path America should follow. When the answers are, to some degree, unknowable, we can at least try to organize intense intellectual effort around the most important questions. The goal of this volume—like RAND’s Geopolitics of AGI project—is to contribute to that task.