A USU graduate student with a background in moral philosophy is preparing to study moral injury in an era of AI-enabled warfare.
 |
| Service members in Army and Navy uniforms work on laptops during a class session. As artificial intelligence becomes part of military medicine and operations, USU is preparing future providers for the ethical questions it raises. (Photo credit: Tom Balfour, USU) |
October 13, 2026 by Hadiyah Brendel
U.S. Air Force 2nd Lt. Raymond Gilmore studied moral philosophy and theology at the Catholic University of America before starting on the path to become a military psychologist at the Uniformed Services University of the Health Sciences (USU). Now a graduate student in USU’s Department of Medical and Clinical Psychology (MPS), he is asking how artificial intelligence on the battlefield could change the moral injuries service members carry home.
Moral injury is the lasting psychological harm, often marked by guilt and shame, that can follow acts that violate a person’s core moral beliefs. As AI enters command-and-control operations, the systems commanders use to direct forces, MPS faculty and graduate students in USU’s School of Medicine are preparing to help service members work through the ethical questions that come with its use in combat, part of USU’s wider focus on artificial intelligence in military medicine.
“I was already tending towards looking at moral issues around the use of drones and remote warfare from a philosophical perspective,” Gilmore explains. “Working with former military and law enforcement personnel, I started asking what issues I’d be most likely to see in the future. The integration of artificial intelligence and machine learning into the military is a front-and-center consideration right now.”
Gilmore’s philosophical background gives him an analytical framework to examine human nature in combat. “It’s kind of incoherent to talk about the use of AI if we’re not even sure what the nature of a human being is and how that nature interacts in an environment of war,” Gilmore says. “These broad questions—what is consciousness? What does it mean for a human to be at war?—are the most important questions we’re digging for answers to around AI right now.”
Calibrating Trust: Automation Bias vs. Algorithm Aversion
A key component of USU’s forward-looking approach is preparing future providers to understand how service members interact with Decision Support Systems (DSS)—algorithmic software designed to analyze battlefield data and provide recommendations, including lethal engagement options.
In high-fidelity simulations featuring active-duty personnel, human interaction with these systems tends to divide into two cognitive groups: automation bias and algorithm aversion.
“Automation bias is relying excessively on recommendations from the machine side as a cognitive shortcut,” says Gilmore. “Algorithm aversion is on the other side—even in the presence of reasonable recommendations or a proven track record, rejecting suggestions simply because ‘I don’t trust machines, I trust humans.’”
 |
| U.S. Air Force 2nd Lt. Raymond Gilmore, a graduate student in the Uniformed Services University’s Department of Medical and Clinical Psychology, studies how artificial intelligence in combat could shape moral injury among service members. (Photo credit: Tom Balfour, USU) |
Because real-world military AI data remains largely classified, Gilmore examines high-fidelity simulations featuring active-duty personnel to observe how these biases manifest under pressure. In these simulated environments, researchers observe how trust fluctuates dynamically based on operational factors. When potential collateral damage to civilians was high, or when fatigue set in, operators exhibited heightened algorithm aversion. Alternately, perceived system competence triggered automation bias, prompting operators to defer critical judgment to software during the simulation.
An underlying tension in the dynamic is Incomplete AI explainability. During the studies, greater trust for algorithmic outputs prevailed when the system provided even a single line of reasoning for its recommendation. However, complex machine-learning models don’t reliably track their own decision process. This disconnect can leave service members to process complex feelings after carrying out machine-assisted choices.
Addressing Moral Injury and "Distantiated Intimacy"
Gilmore noted that while military doctrine still emphasizes keeping a “human in the loop,” meaning a person makes the final decision, operational standards are evolving quickly as autonomous systems emerge around the world.
“Responsibility is huge when looking at moral injury,” Gilmore notes. “It’s a fundamentally different experience to say, ‘I assessed the situation and made the decision to do this,’ versus ‘The system recommended this course of action and I authorized it.’ There’s a different flavor of guilt and shame that comes from diffusing responsibility onto a nonhuman system.”
This shifting dynamic closely connects to how USU researchers examine moral injury in modern warfighters, where accountability and decision-making directly shape long-term psychological health. Gilmore also counters the cultural perception that remote, tech-driven warfare is detached or cold.
“That picture is not at all accurate,” Gilmore stresses. “When you're operating these systems, you experience 'distantiated intimacy'—there's physical distance, but a deep emotional connection as you watch up close over extended periods. The psychological toll is 1,000 percent still there, carrying unique symptom clusters involving guilt and shame.”
For Gilmore and his peers in the MPS program, preparing for the future of military medicine involves forward-looking discussions and adaptable care strategies to ensure providers remain flexible as technology evolves.
“I can't tell the future," Gilmore says. "The best I can do is study hard, engage in this discourse, and prepare myself to be a competent, flexible provider."