Author Description

"Advances in artificial intelligence (AI), machine learning, and robotics have raised the possibility that the profession of arms will soon include integrating human soldiers with AI-enabled machines and applications as part of the collective whole. Machines and software applications enabled by AI are starting to demonstrate capabilities that are relevant to military settings, such as moving autonomously through complex urban traffic and creating startlingly humanlike and interesting derivative works through large language models. However, this does not mean that such developments can be implemented in military settings smoothly. The practice of building cohesive small units is no easy endeavor. The best small units cohere to the point where one soldier recognizes the silhouette and gait of another in the dark of a patrol base in an instant. The best staffs internalize their commander's style and specific needs over time. Integrating humans and machines in military contexts will likely draw from civilian parallels but will also require substantial contextualization. In this report, the authors investigate the kinds of difficulties that the Army might encounter as it attempts to pair humans with AI algorithms to accomplish specific warfighting tasks. They make recommendations to address how these potential obstacles can best be overcome and ensure that the Army effectively creates AI systems that will integrate well with the soldiers who must interact with them."--