Deze toespraak is alleen beschikbaar in het Engels.
Part 1. The opportunities and threats of AI in warfare
Your Eminences,
Your Excellencies,
Distinguished guests,
Leaders from industry,
Ladies and gentlemen,
There are moments that quietly change the way you see the world.
You might expect a moment like that in a place like this: full of history, beauty, and religious significance.
But it happened to me in an underground command post, in the eastern part of Ukraine.
A few weeks ago, I was visiting Major Robert Brovdi, commander of Ukraine’s Unmanned Forces Systems.
Not long ago, Major Brovdi was a clean-shaven businessman, interested in fine arts.
When the large-scale invasion started in 2022, he joined the armed forces.
Now he’s a bearded major in camouflage, running his unit like a business.
His command post is hidden in the oft-attacked region of Dnipro.
Upon entering, I saw a state-of-the-art, modern office with large flat screens lining the walls. On one of these screens, he pointed to a number I had to remember: 140. He didn’t say anything about that number at the time, but he would come back to it at the end of our visit. And I’ll get back to it as well.
He showed me around, and for two hours we spoke about his experiences. What I saw and heard may well be the future of warfare. AI systems processing live drone feeds. Identifying potential targets. Calculating the probability of success.
Every mission generated new data. How many enemy positions had been neutralised. How much ammunition had been used. Which drone had performed best. Which software version had achieved the highest success rate.
Then he showed me another number. At that moment, they estimated that the average cost of eliminating one Russian soldier had fallen to 969 US dollars.
They compared this with another figure: recruiting, training, equipping, and deploying one Russian soldier to the battlefield costs roughly 58,000 dollars.
Ukraine has to constantly calculate whether it can afford to defend itself, and wants to make sure that the enemy can no longer afford this war.
As State Secretary for Defence, I understood exactly why they measure these things. Ukraine cannot afford waste — not of ammunition, not of drones, and certainly not of human lives.
And innovation such as AI allows a smaller democracy to defend itself against a much larger warmongering autocracy that commits war crimes on a daily basis.
Countries like mine are learning from this.
We have to, because we know that Russia’s ambitions go beyond Ukraine.
And what we can see is that Ukraine is using artificial intelligence as one of its military tools to protect humanity.
Its military innovation is crucial and lifesaving – for Ukraine and for Europe.
And yet, as I stood there watching those dashboards... Something didn’t sit right inside.
That dashboard showed the cost of fighting. But not the cost of a human life.
Every symbol on that screen represented a human being. A son or a daughter. A husband or a wife. A father, a mother.
War destroys lives.
So let’s be clear: the Kremlin started this war in and it can end it, today.
But until the Kremlin opts for peace, Ukraine has to defend itself. And it is our duty to help a sovereign state against this ongoing aggression.
Ukraine’s fight is justified and existential.
Even so, the uncomfortable truth is that technology does not only change how we fight. It also changes how we see humanity in the enemy.
It starts with measuring efficiency. Then effectiveness. Then optimisation. Eventually... we begin measuring human beings themselves. Not by their dignity...
But by their utility. Not as persons... But as data. As targets. As variables inside an algorithm.
Of course, the purpose of every military innovation is to be able to fight harder, better, and faster.
What sets AI apart from other military innovations is that it radically changes the speed of warfighting and the involvement of humans in warfighting.
The time between finding targets and striking them, once measured in hours, becomes a matter of mere seconds.
And therein lies the real danger: humans being quietly pushed out of military decision making.
So the question is where, in all this speed, the human being still stands.
That’s where many people’s fear lies.
We restrict warfighting with laws that reflect our values.
The fear is that military AI undermines those values.
Because it’s too fast to govern. Too complicated to regulate. And at the same time, too indispensable.
The challenge is therefore to incorporate AI into the military domain in a legitimate way – without taking away crucial capabilities from our warfighters.
There is a great deal of discussion about the morality of AI.
But let’s look at the morality of the end users of military AI, that of warfighters.
Service members of every rank have to learn how to use military AI to prevent themselves from being overpowered or outsmarted by an enemy that can field greater mass.
But they have to do so according to certain rules.
Humanity is one of the fundamental principles of international humanitarian law, also known as the law of armed conflict.
International humanitarian law therefore seeks, among other things, to regulate warfighting itself, which includes protecting non-combatants.
Even when fighting for a just cause, the line between right and wrong can be thin.
This is where the fundamental principles referred to can offer guidance.
These principles can help service members to understand what is and is not lawful in a given situation.
They can also provide a mandate to act.
And that’s where our challenge lies.
AI is able to surpass humans in speed in every phase of the targeting process.
It enables data-driven decision making on when to act. It strengthens every aspect of defence: from military platforms to logistics, training, planning, and operations.
At its best, AI can help commanders limit collateral damage and protect the men and women in the field.
But we have to recognise the risks.
AI can misidentify targets, amplify errors, blur accountability, and escalate situations beyond human control. Algorithms have no conscience. They do not feel doubt, responsibility, or mercy.
Misalignment, algorithmic hallucinations, cyber vulnerabilities, bias, and data leaks are all legitimate concerns.
Entrusting machines with life-or-death decisions without proper human agency and control is not just risky – it is unacceptable.
And this technology is not confined to state actors. Malign actors are already using AI to threaten our people, our digital infrastructure, and our values.
The standpoint of the Dutch government is that we have to keep human beings in the loop.
At the same time, we saw that Ukrainian troops took back a position using only ground robots and drones earlier this year.
That’s impressive to me.
If working like that can remove the need for Dutch troops to be in the trenches and increase their chances of coming home, then I’m all for it.
But I need a legal framework too.
I don’t want a service member who is using innovative ways to protect us being reprimanded because something wasn’t up to code, ultimately because that code wasn’t clear.
So, military AI is changing warfare.
I’ve laid out the solutions and threats.
Now, we need clear codes and laws that reflect the values that our troops are fighting for.
We need a relationship with military AI that puts humanity first.
Part 2. How should humanity relate to AI?
And that brings me here.
To the Vatican.
Over 2000 years ago, not far from here, the Roman statesman and philosopher Cicero said:
‘Inter arma enim silent leges.’
Translation: ‘In times of war, the law falls silent.’
I am happy to see that this is not true. The lawfulness of modern warfighting is of interest to a wide public. From the barracks to lecture halls, from newspapers to Instagram, and from The Hague to the Vatican.
And the Vatican may be the best place to ask a question that no engineer, no general, and no algorithm can answer:
What is a human being?
Not: what can a human do?
Not: what is a human worth economically?
Not: what military value does a person have?
But: what is a human being?
For Christians, the answer begins on the first page of Scripture:
‘So God created mankind in His own image.’[1]
Everything else follows from that.
Human dignity is not earned, granted, or calculated.
It cannot be measured by productivity, intelligence, beauty, or usefulness.
It exists because every person reflects their Creator.
That includes the child in a Kyiv subway,
the person in uniform serving far from home,
the engineer developing AI—
and the frightened conscript in a trench.
It reminds us: even our enemy is human.
So the question is: how do we preserve and protect our humanity when we use artificial intelligence, especially in the military domain?
We do not defend borders for their own sake.
We defend people.
Every military innovation must serve one purpose: the protection of humanity, even in war.
This is not only a religious idea.
It is the foundation of human rights, international humanitarian law, and military ethics.
But of course, we humans like to – and sometimes need to – go faster, get smarter, do better, get ahead.
Artificial intelligence is extraordinary.
It sees patterns, processes information, predicts outcomes.
But intelligence is not wisdom.
Calculation is not judgement.
Information is not moral understanding.
History is clear: knowledge is power, but that power alone does not prevent cruelty.
Technology alone does not create peace.
That is why we need prudence—prudentia.
AI can suggest what works.
It cannot decide what is right, or true.
That responsibility remains a human one.
If a system fails, it is not the machine that is accountable,
but the people who designed, approved, and deployed it.
This lies at the core of the motto taught at the Dutch Military Academy, where our officers are educated:
I understand my responsibility, I accept the task, and I’m willing to account for it.
That is why human responsibility must remain at the heart of military AI.
Technology may help defend our societies, our values.
Only our humanity tells us why they are worth defending.
As Psalm 139 says:
‘You have searched me and You know me.’
Not analysed.
Not measured.
But known.
AI may know everything about us.
But knowing everything is not the same as truly knowing someone.
Understanding a human being requires more than data.
It requires wisdom.
Conscience.
Compassion.
The future of military AI is therefore not primarily a technological challenge.
It is a moral one.
Not because we fear innovation—
but because every generation must answer the same question:
What kind of people do we wish to become?
Because if we lose sight of humanity along the way,
no success will compensate for that failure.
Because civilisation is not judged by the intelligence of its machines,
but by the wisdom of its people.
And wisdom… can never be artificial.
Part 3 Responsibility and the future we choose
If we see the goal of military AI as protecting humanity, then another question follows:
Who is responsible for shaping this technology?
Today, advanced AI is increasingly concentrated in a few places—above all, the United States and China.
This is not only a technological competition, but a competition of values. Because technology is never neutral. It reflects values, ideas, priorities, and choices—and choices are always moral.
Europe needs to make more use of technologies like AI, cloud and quantum.
We need to understand, produce, and use these technologies ourselves.
It will enable us to better govern and legislate AI.
Humanity has experienced life-changing technology before.
The Industrial Revolution, for example, profoundly changed the way people worked and lived.
American president Theodore Roosevelt saw that to serve both society and giant companies, they both needed a square deal.
This would guarantee equality, stop monopolistic abuse, and shield workers and customers from exploitation.
Roosevelt’s Square Deal did have the effect he hoped for: favourable results for companies and society.
We need a Square Deal for AI, too.
Because it is clear that AI can bring us so much joy, efficiency, safety, and security…
But new trees need to grow on solid ground.
Any AI framework must therefore be grounded in international humanitarian law.
Reflecting on that framework, I am inspired by our Christian humanist tradition…
I am also inspired by my country’s longstanding dedication to the rule of law.
The first work on international maritime law, published in 1609, was written by Dutchman Hugo Grotius.
And one of the first handbooks on the law of armed conflict was written by the Dutch captain Jacobus den Beer Poortugael, in 1872.
This captain wrote that it is our duty to do to one another as much good as possible in times of peace… and as little harm as possible in times of war.
These values still ring true, and our laws keep being updated to reflect those values.
We are the only country in the world that has enshrined the development of the international legal order in its constitution.
That is why the Netherlands initiated the first REAIM Summit in The Hague in 2023.
REAIM stands for Responsible AI in the Military Domain.
The Netherlands has co‑hosted every summit, and we will co-host again, next year in Kenya.
We are building momentum so that we can move from theory to practice, from principles to workable criteria for our warfighters. They need their AI to be trustworthy, lawful, and understandable, from warfighter to developer, and trusted between states.
While this is still a work in progress, the Netherlands’ core message is that people must always be at the centre.
We are committed to developing, acquiring, and deploying military AI responsibly. So that the men and women who command these systems can trust them, and answer for them. Responsible AI keeps human judgement where it belongs: with humans.
When technology becomes foundational, accountability must follow – not just locally or regionally, but globally.
Innovation needs freedom – but also clear limits on what is allowed and what is not.
Democracy needs accountability.
Security needs cooperation.
Our task is to preserve all three.
Because ethics without innovation becomes irrelevant.
Innovation without ethics becomes dangerous.
We need both.
Europe must therefore invest—
in research,
in talent,
in companies,
and in its own technological capabilities.
To become a stronger Europe, and a stronger ally within NATO.
Finally, to the innovators in this room:
You are not just building technology.
You are shaping the future of warfare.
That responsibility calls for humanity.
Because every breakthrough raises three questions:
Can we build it?
Should we build it?
And what are we trying to protect?
Military innovation should not only make military operations more effective.
It also has to take into account what it’s like for the service members using it.
Men and women in uniform who serve polities that are committed to the rule of law are required to act lawfully and use force to protect others while being scrutinised by the media, the political establishment, and society.
And it must help make peace more secure.
Technology will shape the future.
But values will determine whether that future remains worthy of our humanity.
Part 4. The human face of the future.
Ladies and gentlemen,
Allow me to take you back underground, to that command post in Ukraine.
At the end of my visit, the commander showed me the screen again. At the beginning, he had told me to remember the number on that screen: 140. The number had changed—from 140 to 211. During our conversation, 71 Russian troops had been killed.
If Putin had called for an end to the war, every Ukrainian would put have put down their weapon and unplugged those screens.
But he hadn’t.
So the screens kept glowing.
The algorithms kept calculating.
The software kept learning.
Those two hours underground changed me.
As a person.
As State Secretary for Defence.
On that day in Dnipro, I got a glimpse of how indispensable AI is for warfighting…
…but also of how challenging it will be to make sure that humanity stays at the core of military innovation.
So that we protect not just our territory, but also the values that make us human.
Military AI will continue to improve.
It will become faster, more accurate, more powerful.
The real question is whether we will improve as well.
Not only in intelligence, but in wisdom.
Not only in capability, but in character.
That has always been the true challenge of every technological revolution:
not whether we can build extraordinary machines,
but whether we remain capable of raising extraordinary human beings.
As State Secretary for Defence, I know freedom needs soft power and hard power.
Ukraine reminds us of that every day.
That is why we must continue to innovate—confidently, responsibly, morally. Together.
AI can help defend freedom.
It can save lives.
It can protect civilians.
But we must never confuse technological progress with moral progress.
So let us build.
Let us innovate.
Let us cooperate.
But above all, let us remember what we are trying to protect:
Not algorithms.
Not markets.
Not even states.
But people.
Thank you.