During most of my Army career, commanders confronted a terrible responsibility. Intelligence officers assembled the picture. Staffs developed options. Commanders weighed necessity against risk to troops and civilians. An accountable human being made the final decision.

Artificial intelligence is changing that process.

During Operation Epic Fury against Iran this year, the Pentagon’s Maven Smart System operated inside the American targeting architecture. Pentagon AI chief Cameron Stanley testified the system helped strike more than 13,000 targets in 38 days, compressing targeting from days to seconds. CENTCOM’s Adm. Brad Cooper affirmed that human commanders retained formal authority to authorize attacks.

But that reassuring phrase, humans remained in the loop, does not settle the matter.

If artificial intelligence finds the target, organizes intelligence, sets its priority, recommends the weapon and hands a commander a nearly completed course of action, how much of the decision remains genuinely human?

For Christians, that is not merely a technology question. It is a Just War question.

I warned in “The New AI Cold War” that warfare was becoming a contest of algorithms, data, and machine speed. My worry now is that the Pentagon’s enthusiasm for decision speed could push human judgment into a secondary role.

The Pentagon’s pursuit of machine-speed decision advantage is creating operational pressures that could erode the deliberation, independent judgment, and meaningful human responsibility on which Just War reasoning depends.

That does not mean the Pentagon has abandoned Just War principles. The technologies it embraces simply make them harder to practice.

The revolution extends well beyond autonomous weapons. The Army’s Next Generation Command and Control architecture is now “ready to scale,” and the Marine Corps’ Project Dynamis has already demonstrated machine-to-machine targeting in exercises.

The common thread is not simply autonomy. It is dependence.

That dependence raises a question more consequential than whether humans technically remain “in the loop”: Can Just War reasoning survive when the machine sets the tempo and frames the commander’s choices?

Just War reasoning, developed through Augustine, Aquinas, and later moral philosophers, begins before the first shot: is there a just cause, legitimate authority, necessity, proportionality, exhausted alternatives, and reasonable prospect of success?

Those are moral judgments, not optimization problems.

Imagine the decision chain: an AI system flags adversary movement, estimates an attack is imminent, and calculates that delay reduces the chance of success.

The president still decides.

By the time that decision reaches him, AI may already have framed the threat and calculated the cost of waiting. The human retains authority to decide, but the machine has already shaped the terms of it.

A machine does not need the authority to start a war in order to make war appear unavoidable.

That is where machine speed could begin quietly eroding jus ad bellum, the moral reasoning governing whether war should begin at all.

The same problem continues once the shooting starts.

Christian Just War principles require discrimination between combatants and civilians, proportionality, and accountability. I argued in “The New AI Cold War” that human discernment must remain central, since machines can calculate what is possible but not what is just.

Maven shows why that distinction has become urgent. Researchers have documented how the system aggregates intelligence and ranks courses of action before a human ever says yes or no. It operates upstream of the lethal decision, and the Red Cross warns such speed can lead a human simply to rubber-stamp the output.

That means merely keeping a person somewhere “in the loop” is no longer a sufficient moral standard.

There is a difference between human presence and human judgment.

A commander who has only seconds to approve what a machine assembled in milliseconds may technically remain in the loop. He may no longer meaningfully command it.

That is especially troubling because Just War reasoning sometimes demands what an optimization system is built to overcome: delay, doubt, verification, restraint, even accepting tactical disadvantage rather than commit an unjust act.

AI rewards speed and efficiency. Just War asks a different question, not what is fastest, but what is right, and Christian ethics has always been willing to trade speed for patience when the stakes demand it.

The unresolved strike on the Shajareh Tayyebeh girls’ school in Minab, Iran, illustrates the danger without proving an AI failure. Senator Jack Reed (D-R.I.) and other senators pressed the Pentagon over reports that an analyst flagged the site’s changed status years earlier, information that never reached the targeting database. There is no public proof Maven caused the strike.

But the lesson is still sobering: Machine speed cannot make bad data true.

Nor does a human approval click turn an inadequately examined machine recommendation into moral judgment.

None of this means America should abandon military AI. China, Russia, and other adversaries are pursuing military AI, leaving Washington little option but to compete. Used wisely, AI could strengthen discrimination and proportionality.

The choice is not between AI and morality. It is whether AI will remain subordinate to morality.

America must win the AI competition without letting it redefine what America considers morally permissible. Pentagon policy already requires commanders to exercise “appropriate levels of human judgment over the use of force.” That standard must be real, not ceremonial: traceable accountability for what the machine recommended, and the ability to fight when the AI fails.

A military that can fight brilliantly at machine speed but cannot fight when the machine goes dark has merely exchanged one vulnerability for another.

Scripture recognizes government’s legitimate authority to use force for justice. Peter tells believers to submit to institutions that punish evildoers and praise those who do good (1 Peter 2:13-14). That authority belongs to accountable humans, not to the sword itself.

Neither does the algorithm attached to it.

Genesis gives the deeper reason: God made man in His own image (Genesis 9:6). A machine bears no such image. It has no conscience, no moral agency, and will never answer before God for a life taken. Humans will.

The worry, then, is not that machines are becoming persons. It is almost the opposite.

Machines are not persons, yet we keep handing them functions that require the judgment of persons.

Artificial intelligence may help American forces see farther, decide faster, and strike more precisely than ever. We should welcome that when it protects the innocent and defends the nation.

But efficiency cannot become sovereignty. The commander cannot become the machine’s assistant, and moral judgment cannot become the slow part of warfare that we automate away.

America may need machines to fight tomorrow’s wars. But if we allow the machine to define what is necessary, proportional, and just, we may win the battle at machine speed while surrendering the moral judgment that makes a war worth winning.