Insights from the Development and Play of Tsunami

Artificial intelligence (AI) is an increasingly important national security topic, with both its short- and long-term implications for all elements of U.S. society becoming important to analyze. In particular, the topic of artificial general intelligence (AGI)—defined as extremely capable AI with equivalent capabilities to humans across most, if not all, domains—has become important as this technology moves from fiction to an uncertain, but technically credible, possibility that is actively being pursued by leading AI labs.

This paper describes the game Tsunami, which was designed to explore the impact of the development and deployment of AGI on state-state relationships and geopolitics. Tsunami provides insights into the decisionmaking processes of national leaders grappling with the development and integration of AGI into their societies amid economic, military, and cultural challenges. By simulating the intricate interplay between advancing technology and global politics, Tsunami seeks to explore the uncertainties that AGI might bring about and to begin illuminating the dynamics that will underpin AGI’s impact on geopolitics.

This paper is intended to present the lessons that emerged from the development and play of Tsunami, including lessons that arose from the game itself regarding the potential behavior of nations during the development and deployment of AGI and some broader lessons on the design of the game and their application to policy issues, such as AGI.