The integration of machine learning into military decisionmaking is widely seen as critical for the United States to retain its military dominance in the 21st century. Advances in machine learning have the potential to dramatically change the character of warfare by enhancing the speed, precision, and efficacy of decisionmaking across the national security enterprise. This report explores how machine learning can be leveraged to enable military decisionmaking at the operational level of competition and conflict as part of a collaboration between machine learning tools and human analysts. The authors present a case study based on a machine learning-based analysis of real-world data about the conflict in eastern Ukraine prior to Russia's 2022 invasion. This case study places the reader in a commander's shoes, tasked with making decisions about the best types of support to provide Ukrainian forces to achieve shared objectives. This analysis demonstrates that machine learning can improve efficiency by helping human analysts leverage massive data sets that would be impractical for humans alone to examine. The authors found that machine learning has great potential to enable military decisionmaking at the operational level of war but only when paired with human analysts who possess detailed understanding of the context behind a given problem.

Chapter One: Machine Learning as a System -- Chapter Two: Demonstrating the System at Work: Machine Learning and the Conflict in Ukraine -- Chapter Three: Strengths and Limitations of Machine Learning -- Chapter Five: Implications for the U.S. Army -- Appendix: Training a Machine to Read the News.