Can a Machine Be Permitted to Decide Who Dies?

Some decisions can be accelerated without changing what they are. The decision to take a human life is not one of them.

This week the United Nations Secretary-General and the president of the International Committee of the Red Cross renewed their call for legally binding prohibitions and restrictions on autonomous weapons. The appeal comes before November’s review conference for the Convention on Certain Conventional Weapons.

Weapons with autonomous functions already exist. Once activated, some systems can select and engage targets using sensor data without further human intervention. Reports that fully autonomous systems are independently targeting people in current wars remain unconfirmed, but the technological direction is no longer hypothetical.

The immediate danger is civilian harm in environments machines cannot fully understand. A combatant may surrender. A wounded person may become protected. A child may enter a space the system classified seconds earlier.

Beneath that danger lies the deeper question of agency.

A machine can execute a decision. It cannot answer for one.

Scripture

Human beings bear the image of God. Their lives cannot be reduced to target profiles, and their moral agency cannot be transferred to code.

Genesis 1:26–27

A machine cannot understand the worth of the person before it. It cannot exercise mercy, feel the gravity of killing, confess error, repent or stand before a court. That does not make human soldiers infallible. It makes human accountability indispensable.

Technology can assist identification, defence and precision. But efficiency is not the highest good in war. International humanitarian law already places duties on people, not weapons. Delegating the final judgment does not remove those duties; it risks making their bearer harder to find.

Responsibility begins before deployment. Engineers, commanders, ministers and states cannot divide a lethal system into so many small decisions that nobody remains answerable for the whole.

Christians should resist both technological panic and technological surrender. The question is not whether machines are evil. It is whether people may design a system that places lethal judgment beyond meaningful human control and then call the resulting distance responsibility.

States still have time to draw a line. Waiting for an atrocity would mean asking a machine-made grave to teach us what human dignity had already said.