1. The term “autonomy” refers to the ability of a machine to perform a task without human input. This is conceptually different from two other terms: “automatic” (systems that have very simple, mechanical responses to environmental inputs, e.g., trip wires and mines) and “automated” (more complex, rule-based systems, e.g., self-driving cars and modern programmable thermostats). There is no internationally agreed-upon definition for “autonomous weapon.” This study adopts the definition provided in U.S. DODD 3000.09. An autonomous weapon system is a weapon system that, once activated, can select and engage targets without further intervention by a human operator. This is conceptually different from “human-supervised autonomous weapon system” (an autonomous weapon system designed to provide human operators with the ability to intervene and terminate engagements) and “semi-autonomous weapon system (a weapon system that, once activated, is intended to only engage individual targets or specific target groups that have been selected by a human operator). Paul Scharre and Michael Horowitz, “An Introduction to Autonomy in Weapon Systems,” Project on Ethical Autonomy Working Paper, Center for a New American Security, February 2015, https://s3.us-east-1.amazonaws.com/files.cnas.org/documents/Ethical-Autonomy-Working-Paper_021015_v02.pdf?mtime=20160906082257; United States Department of Defense, Autonomy in Weapon Systems, DOD Directive 3000.09 (Washington, DC: Department of Defense, 2012).
2. Congressional Research Service, Artificial Intelligence and National Security, R45178 (2020).
3. United States Department of Defense, Summary of the 2018 Department of Defense Artificial Intelligence Strategy (Washington, DC: Department of Defense, 2019), https://media.defense.gov/2019/Feb/12/2002088963/-1/-1/1/SUMMARY-OF-DOD-AI-STRATEGY.PDF.
4. All current AI systems fall under the Narrow AI category, which refers to algorithms that address specific tasks. The most prevalent approach to Narrow AI is machine learning, which involves statistical algorithms that replicate human cognitive tasks by driving their own procedures through analysis of large training data sets. It will take much longer to develop General AI, which refers to more complicated systems capable of human-level intelligence across a broad arrange of tasks. Congressional Research Service, Artificial Intelligence and National Security.
5. China State Council, “Guowuyuan Guanyu Yinfa Xin Yidai Rengong Zhineng Fazhan Guihua de Tongzhi” 国务院关于印发新一代人工智能发展规划的通知 [China’s Next Generation Artificial Intelligence Development Plan (2017)], July 8, 2017, http://www.gov.cn/zhengce/content/2017-07/20/content_5211996.htm.