On the Geopolitics of AGI
Insights from a RAND initiative to explore the strategic implications of AGI
by Jim Mitre and Joel B. Predd
A decade ago, few believed that artificial general intelligence (AGI)—human-level or superhuman-level cognition across a wide variety of tasks—would emerge in our lifetime. Today, policymakers and executives worldwide are confronting the possibility that AI systems could soon match or exceed human performance in nearly all economically and militarily significant domains. Whether leading AI companies cross the unknown, potentially unknowable threshold to AGI today or tomorrow, we will live for the foreseeable future in a world where increasingly advanced AI underpins transformational changes to economies, militaries, and societies. Moreover, this prospect of technological change coincides with a period of profound shifts in geopolitics and global security, as the postwar consensus erodes and the international system is once again characterized by explicit great-power competition.
At RAND, we are committed to helping decisionmakers navigate the possibility that leading AI companies are nearing the development of AGI and the transformational changes from advanced AI that disrupt geopolitics and national security in a world in transition. We launched the Geopolitics of AGI Initiative in December 2023 to focus our efforts, bringing together technologists, strategists, diplomats, economists, philosophers, regional experts, and more to address one key question: How should the U.S. government respond to the potential emergence of AGI?
The scope of our initiative is broad and diverse, and in the best of RAND traditions, it is fueled by debate and disagreement. We acknowledge that timelines, technical paradigms, and implications are uncertain, but we accept AI that substitutes for or exceeds human capabilities as a technically credible possibility. We believe that the alternative assumption exposes the United States and humanity to an unacceptable risk of being intellectually bankrupt the moment AGI arrives, whenever and however it arrives, if it arrives. Broadly, our research aims to identify different AGI-enabled capabilities; explore their operational, strategic, and global implications; and analyze policies to address the risks and opportunities presented by these technologies.
Today, we launch the Geopolitics of AGI Substack to share the fruits of our initiative. We remain committed to RAND’s well-known, long-standing principles of objectivity, quality, and analytical rigor, so the Substack is a complement to (not a substitute for) the longer-form essays and research we will commission and publish. This Substack is best viewed as an experiment in sharing insights, hypotheses, and ideas on timelines and formats that contribute to an increasingly active and urgent public debate.
We hope that you will join us in making the Geopolitics of AGI Substack a success by signing up here. We welcome your feedback and reactions.