This article, written by Michael C. Horowitz and Erik Lin-Greenberg, explores how AI affects decision-making during times of crisis. Horowitz and Lin-Greenberg use two original survey samples to test out how the US government acts based on AI versus human analysis and how the government would respond to an adversary that utilized AI. The authors delve into contrasting scenarios, particularly the effects of personal, friendly AI use in the military versus the AI use of adversaries, particularly in hostile ways. The authors find that experts are less likely to use military force when information is received from AI rather than human analysis. This can be attributed to the fact that it is difficult to assess the accuracy of the information received and current trust for AI is still quite limited. In addition, experts are more likely to use military force in retaliation when AI kills an American soldier rather than a human adversary. The findings of this paper highlight the importance of human decisions for international security in the wake of new emerging technologies. This article prompts a deeper reflection on the relationship between humans and AI when it comes to military decision-making. Horowitz and Lin-Greenberg successfully articulate this nuanced relationship, which is essential for policymakers and officials to understand as the climate of emerging technology and national security challenges are constantly evolving.
