Published March 8, 2026
The old slogan “guns don’t kill people; people do” is frequently invoked in debates over gun control. Though glib and often abused, it points to a real moral issue: human agency. But when applied to modern questions of warfare involving AI, it needs to be updated — AI should not kill enemies; soldiers should. The question facing the United States is not whether AI will be used in war — it already is — but whether we will preserve the moral architecture that rightly places responsibility for killing in human hands.
Recent debates between AI developers and the Pentagon expose a deep moral question: Who bears responsibility when lethal force is used? Who ought to possess ultimate authority over life-and-death decisions: human agents or autonomous machines? In the end, who — or what — pulls the trigger? At present, thankfully, AI companies such as Anthropic are insisting that their technology not be used autonomously — despite mounting pressure from the Pentagon, which is now taking steps to part ways with the company. As war begins between the U.S. and Iran, these questions are especially pressing, and they will remain relevant in all future armed conflicts.
On ethical grounds, AI companies should insist that AI-enabled weaponry preserve human agency in lethal decision-making. AI can identify legitimate targets, assess enemy patterns, and synthesize intelligence at speeds no soldier could match. But a human being must retain responsibility for authorizing the use of deadly force. At present, this remains the stated policy of the U.S. military. But as AI technology continues to rapidly evolve, we should reinforce our commitment to prohibiting AI-activated killing.
Ironically, one way we preserve our humanity in war is by acknowledging that the taking of human life — when tragically necessary — must be carried out by human agents acting under just and lawful conditions. Christian Scripture explains this grim moral responsibility: “Whoever sheds the blood of man, by man shall his blood be shed, for God made man in his own image” (Genesis 9:6). This is not a license for bloodlust. It restrains warfare by grounding the use of force in proportionality and responsibility. It also reminds us that human life bears a unique moral weight, because human beings are made in God’s likeness (Genesis 1:26–27). The one who takes life bears moral responsibility; the one whose life is taken bears moral culpability. That symmetry makes sense only if we believe there are objective moral conditions under which lethal force may be tragically justified — and that humans, not machines, are accountable to them.
The Christian just war tradition rests on a deeper anthropology helpful for the debate at hand. It insists that even an enemy is a person — a morally responsible agent who can be held accountable. To kill “justly” is to recognize that people are not mere datasets. When we shift the decision to kill to an algorithm, we do more than increase the chance of victory in war; we erode the moral relationship between human beings. That moral relationship is what helps us keep our humanity. Recognizing that war requires the killing of human beings helps discipline the moral imagination — not to lust for war, but to limit its frequency and restrain its means.
The Pentagon’s operational logic — speed, lethality, scale, survivability — is understandable. No serious ethicist denies the permissibility of AI-enabled assistance. In high-velocity conflicts, milliseconds matter. But the move from “assistance” to “authorization” is a tectonic moral shift. It changes who holds moral responsibility when the shot is fired. And the answer should not be machines.
There is also irony here. A system that never feels the weight of killing can never be humane in the way that matters. Taking a life carries an enormous psychological burden, but that is precisely what a machine cannot feel — and what soldiers most need to remain humane. Outsourcing the decision to kill can also deaden the conscience by cheapening how we evaluate life’s value. There are studies, for example, that show drone operators — who dehumanize enemy combatants with remotely authorized killing — have suffered real moral injury. There’s a corollary here: drones currently rely on human-activated triggers, and we should insist that future machines do the same. Mechanized killing has bodily consequences for the combatant — their death — and a moral consequence for the warrior — a wounded conscience. It is not as efficient or morally sanitized as we may think.
The proper role of AI in war is to augment human judgment, not replace it. The moment we let algorithms decide when a human being dies, we move beyond the tragic necessity of force into the bureaucratization of killing. A nonnegotiable moral standard is that AI must never be granted full autonomy over the decision to kill.
In short: A lawful human enemy is a person, not simply a dataset.
Andrew T. Walker, Ph.D., is a Fellow in the Evangelicals and Civic Life Program at The Ethics and Public Policy Center. He is also the Dean of the School of Theology and Professor of Christian Ethics and Public Theology at The Southern Baptist Theological Seminary in Louisville, Kentucky, and the founding Managing Editor of WORLD Opinions. He co-hosts The Bully Pulpit, a weekly podcast exploring faith, culture, and politics from a conservative Christian perspective.
