Heeding the Risks of Geopolitical Instability in a Race to Artificial General Intelligence

As artificial intelligence (AI) rapidly advances, many AI experts predict that the first state to develop artificial general intelligence (AGI) that can perform a wide range of tasks better than humans will gain huge advantages in military and economic power. If U.S. and Chinese leaders believe that losing the race to AGI would pose a dire threat to their nations, how will they respond if their strategic competitor appears poised to win it? Or, how will they respond if their state successfully develops AGI and then faces challenges to its newly achieved technological dominance?

The author of this paper presents a typology of potential strategic responses, focusing on preventive actions that states might take to undermine an opponent's AI development efforts, and draws on the history of geopolitical power shifts and nuclear proliferation to identify key factors that are likely to affect whether national leaders will decide to launch preventive attacks against rival AI programs. Uncertainties about the potential characteristics and implications of AGI might make pressures for preventive action especially powerful but might also discourage leaders from taking great risks when the magnitude and proximity of the danger are unclear. This analysis suggests that strategists and decisionmakers should seriously consider how incentives for preventive action might make the period of transition before and after the emergence of AGI geopolitically fraught. The assessments presented in this paper can provide a starting point for making policy choices to manage the resulting risks of international instability.