AI, Autonomous Weapons and Human Rights: When Machines Decide Who Lives and Dies

Before the code, there was conscience.

A hand once raised a prayer,

A hand once held a blade.

Now circuits learn to calculate

The choices humans made.

A face becomes a number,

A heartbeat becomes a sign.

Prayer speaks of human dignity;

The machine reads a design.

Between the soul and silicon,

Between mercy and command,

A question older than our wars

Now waits inside our hands.

Humanity has always created tools to extend its power. We have built machines that can fly faster, travel further and see beyond the limits of our own eyes. Artificial intelligence is now extending that power into warfare.

But there is a boundary that should make every society pause.

What happens to human dignity when we allow machines to make decisions about life and death?

This is no longer merely a question for computer scientists or military planners. It is a question for philosophers, religious leaders, governments, soldiers, parents and ordinary citizens.

The United Nations and the International Committee of the Red Cross are calling for stronger international rules on autonomous weapons, including legally binding measures. In August 2026, UN Secretary-General António Guterres and ICRC President Mirjana Spoljaric renewed their call, warning about the growing danger of machines autonomously targeting human beings.

The machine that does not understand life

There is an important difference between a weapon controlled by a person and one that can independently identify, select and attack a target.

AI can assist a human being. It can analyse enormous quantities of information and help identify objects or potential threats. More autonomous systems can take on critical functions in finding and engaging targets.

The technology may be remarkably sophisticated.

But sophistication is not conscience.

A machine may recognise a human shape. It may calculate distance, movement and probability. It may determine that something fits the characteristics of a target.

But does it understand fear?

Does it understand a mother's grief? Does it understand innocence? Does it understand mercy, forgiveness or the value of a life that has taken decades to build?

Recognising a human shape is not the same as understanding a human being.

That distinction lies at the heart of the debate.

Religion and the meaning of human life

Long before computers existed, human societies wrestled with the morality of killing.

Religious and spiritual traditions differ profoundly from one another, but many have developed rich teachings about the value of human life, compassion, responsibility, mercy and the moral consequences of violence.

In Christianity, Islam, Judaism, Hinduism, Buddhism and many other traditions, human life is treated as something carrying profound moral significance, although the reasoning and teachings are not identical.

Even traditions that have grappled with the ethics of warfare have asked difficult questions about necessity, restraint, responsibility and the treatment of others.

That history matters.

Because the question before us is not simply whether a machine can kill accurately.

It is whether a machine should ever be entrusted with the moral decision to kill a human being.

Can a machine possess a conscience?

It can be programmed with rules. It can be trained on enormous quantities of information. It can respond to changing circumstances.

But conscience is more than processing information.

It involves responsibility. It involves judgement. It involves the uncomfortable knowledge that another person is not merely an object standing between us and an objective.

If a human being makes a decision to take a life, that person may have to live with the moral and legal consequences.

What happens when the decision belongs partly to an algorithm?

Culture, history and the human experience

Every culture has developed its own ways of understanding death, mourning and conflict.

Across generations, people have created ceremonies for the dead, stories about courage and sacrifice, laws governing warfare and traditions intended to restrain cruelty.

Even in war, humanity has repeatedly tried to establish boundaries.

Now imagine a battlefield where an algorithm identifies a possible target thousands of miles away.

There may be no face-to-face confrontation. No eye contact. No human voice. No moment in which one person looks at another and understands that the figure in front of them is a living human being.

Perhaps that distance could reduce the emotional burden of warfare.

But perhaps it could also make killing easier.

Technology can reduce suffering. It can improve precision and protect human lives. That should not be ignored.

The danger is not technology itself.

The danger is allowing technological capability to move faster than our moral judgement.

When the battlefield becomes a calculation

A machine sees a battlefield differently from a human being.

Coordinates.

Probabilities.

Heat signatures.

Movement.

Targets.

But a human being might see something else.

A child.

Someone's father.

Someone's daughter.

A friend.

A frightened civilian running towards safety.

A person carrying memories, hopes and people who love them.

The machine may calculate the likelihood of a threat.

But who calculates the value of a life?

Who decides when uncertainty is too great?

And when the machine is wrong, who carries the responsibility?

Who should have the final word over life and death?

Human rights and meaningful human control

This is why autonomous weapons are fundamentally a human-rights issue.

The concern is not simply whether a weapon is technically accurate. International humanitarian law already requires parties to armed conflict to respect important rules concerning civilians and the conduct of hostilities.

The deeper question is whether human beings remain meaningfully responsible for the use of force.

The idea of meaningful human control is therefore crucial.

The ICRC has called for legally binding rules that would prohibit certain autonomous weapons and strictly regulate others. Its position includes opposition to unpredictable autonomous weapons and to systems designed or used to apply force directly against people, alongside requirements for human supervision and the ability to intervene where appropriate.

The UN has likewise called for a legally binding instrument and has stressed that responsibility for decisions about the use of force must remain with humans.

The practical question is uncomfortable.

If an autonomous weapon kills the wrong person, who is responsible?

The soldier who deployed it?

The commander?

The programmer?

The manufacturer?

The government?

Or does responsibility become so divided between human beings and machines that nobody truly owns the decision?

Human rights cannot function properly in a world where responsibility disappears into software.

When efficiency becomes the enemy of wisdom

One of the great temptations of artificial intelligence is efficiency.

A machine can analyse information in milliseconds.

But speed is not the same as wisdom.

Human judgement is often slow precisely because human beings must consider context. We hesitate because another person's life matters.

An algorithm might be able to calculate probabilities with astonishing speed, but morality is not simply a mathematical equation.

Intelligence can process.

Judgement must understand.

And wisdom sometimes means refusing to act.

This raises a much larger question.

If we can give machines the power to make decisions, does that mean we should?

Human beings have always created tools to extend the hand. But there are some responsibilities we should not extend beyond the human conscience.

The concern is not that AI will suddenly become evil.

The concern is that human beings may become too willing to hand over responsibilities that should remain ours.

The boundary we choose

Autonomous weapons force us to decide what kind of future we want.

A future in which machines assist human beings while humans retain meaningful responsibility is one possibility.

A future in which machines independently determine who is targeted and killed is another.

The difference between the two is not merely technical.

It is moral.

Our ancestors asked whether there were limits to what could be done in war. We are now being asked whether there should be limits to what machines are allowed to decide.

Perhaps the greatest measure of our intelligence will not be how much power we can give our machines.

Perhaps it will be whether we have enough wisdom to know which powers should remain human.

And so we return to the hand.

The hand that builds.

The hand that heals.

The hand that prays.

The hand that can destroy.

Beside it stands the machine — faster, stronger and increasingly capable.

But the final choice must remain ours.

Let the machine calculate,

Let the machine assist,

Let human knowledge reach

Beyond what eyes can see.

But keep the hand upon the conscience,

Keep compassion in the frame.

For no child should become a number,

And no life should be a name

That disappears into a programme,

A target, code or chart.

The future we give our children

Must still have a human heart.

Perhaps that is the real test of intelligence: not whether we can teach machines how to make the ultimate decision, but whether humanity is wise enough to keep that decision where it belongs — with the human conscience