They become nothing more than a rubber stamp for a machine
From destroyer decks in distant seas, missiles reach neighborhoods before accountability can follow — this is the shape of modern war. Researcher Lauren Gould has spent over a decade tracing how Western militaries have traded soldiers on the ground for algorithms in the cloud, and what that trade costs the people below. The US-Israel air campaign against Iran, the algorithmic targeting in Gaza, the secret bombing of Hawija — each is a chapter in the same story: that when warfare becomes remote enough, the dead become data, and the data is often wrong. What is being perfected, Gould warns, is not precision but scale — and scale, in urban warfare, kills civilians nine times out of ten.
- Since February, the US and Israel have struck over 13,000 targets in Iran — including a primary school where more than 175 children and teachers were killed on a single day.
- In Gaza, an Israeli AI system flagged 37,000 people as high-risk; operators had 20 seconds per designation, reducing life-or-death decisions to a rubber stamp on an algorithm no one can verify.
- The Dutch government concealed its role in the Hawija bombing — which killed at least 85 civilians — for four years, and communities who had trusted Western precision felt not just grief but betrayal when the truth emerged.
- Surveillance systems create what Gould calls 'psychic imprisonment': civilians cannot determine what behavior is safe, because the criteria for being marked dangerous are invisible and arbitrary.
- Tech giants supply the cloud infrastructure enabling this system, and when public pressure forces one company to step back, another steps in — the machinery never actually stops.
- Gould argues these wars are no longer fought to achieve political outcomes — the kill chain has become an end in itself, sustained by economic interests and the commercial appetite of weapons and technology industries.
The missiles arrive before anyone sees them coming. A destroyer somewhere in the Persian Gulf launches a cruise missile, and by the time it lands, the operator is already authorizing the next strike. Since February, the United States and Israel have struck more than 13,000 targets in Iran — over 11,000 munitions in the first 16 days alone. Western publics see the launches. They do not see the school struck on February 28, where more than 175 children and teachers were killed.
Lauren Gould, a conflict studies researcher at Utrecht University, has spent over a decade studying what she calls 'remoting' — the deliberate shift toward drones, satellites, and intercepted data in place of soldiers on the ground. The logic emerged from Afghanistan and Iraq: fewer military casualties, sustained public support. The flaw is structural. Fewer boots on the ground means less human intelligence. Decisions become only as good as the data — and remote data is often deeply unreliable.
In 2015, a Dutch F-16 struck what was believed to be an ISIS munitions depot in Hawija, Iraq. The explosion destroyed a neighborhood and killed at least 85 civilians. The Dutch government kept it secret for four years. When the truth emerged, governments accepted responsibility but denied liability, and no individual compensation was ever paid. Gould found something striking in her fieldwork there: locals had believed in the precision narrative. When the strike killed so many civilians, the only explanation that made sense to them was that it must have been intentional. The harm extended far beyond the initial death toll — lost income, displacement, fractured communities, terror that continued long after the bombs stopped.
Now artificial intelligence has accelerated the system. In Gaza, an Israeli military algorithm flagged approximately 37,000 people as high-risk. Operators had 20 seconds to verify each designation. Told the system was 90 percent accurate — a claim with no verified basis — they became rubber stamps, sometimes checking only whether a proposed target was male. If male, proceed. The machine absorbed their sense of responsibility.
The surveillance layer compounds the harm in a different way. In Gaza, algorithms were reportedly trained on behaviors like changing phones frequently or appearing in a WhatsApp group with someone on a watchlist — ordinary survival behaviors in a war zone. People cannot know what might classify them as dangerous. Gould calls this psychic imprisonment: the freezing of daily life under invisible criteria, the inability to determine what is safe. The harm does not end when the bombing stops.
Amazon, Google, and Microsoft all provide the infrastructure that makes this possible. When public pressure mounts, companies temporarily withdraw — Microsoft did so after its Azure services in Gaza became public — but the work continues through other contracts, other companies. The system never actually ends.
Gould's deeper argument is this: these wars are not being fought to win. Political objectives have given way to economic ones — access to markets, resources, the perpetuation of the military apparatus itself. The kill chain has become an end in itself. To change that requires naming what war actually is: not incidents that went wrong, but weapon systems deliberately engineered to increase the speed and scale of killing. That is strategy. And that is a very different conversation.
The missiles come from so far away that no one sees who fires them. A US destroyer sits somewhere in the Persian Gulf or the Arabian Sea, and a cruise missile launches into the sky. By the time it reaches its target, the person who authorized the strike is already looking at the next one. This is modern warfare: algorithmic, remote, and almost entirely invisible to the people it kills.
Since February, the United States and Israel have launched an air campaign against Iran that has struck more than 13,000 targets. In just the first 16 days, according to the Royal United Services Institute, over 11,000 munitions were deployed. The scale is staggering. But the Western public sees only what militaries choose to show: cruise missiles launching from ship decks, distant explosions lighting up night skies. What remains hidden is the ground-level reality—the neighborhoods flattened, the schools destroyed, the bodies counted by people who knew the dead.
On February 28, a strike hit an Iranian primary school. More than 175 children and teachers were killed. Lauren Gould, an associate professor of conflict studies at Utrecht University who has spent over a decade studying how Western militaries wage war from a distance, calls this the consequence of what she terms "remoting"—the deliberate shift toward waging war with fewer soldiers on the ground and greater reliance on drones, satellites, and intercepted communications. The logic seemed sound after Afghanistan and Iraq: keep fighting, but reduce military casualties at home, maintain public support. The problem is fundamental: if you have fewer people on the ground, you lose human intelligence. Your decisions become only as good as your data. And remote data is often not very good at all.
In June 2015, a Dutch F-16 struck what was believed to be an ISIS munitions depot in Hawija, a city in northern Iraq. The explosion destroyed an entire neighborhood and killed at least 85 civilians—the second-deadliest single attack of the international coalition's campaign against ISIS. The Dutch government kept it secret for four years. When the truth emerged, successive governments accepted responsibility but denied liability, arguing the target was legitimate and blaming Americans for providing faulty information about the explosives present. No individual compensation was ever paid. Gould conducted fieldwork in Hawija and found something revealing: local civilians had believed in the precision narrative. They had accepted that Western technology was advanced enough to distinguish combatants from civilians. When a strike killed large numbers of civilians, they felt doubly betrayed. The only explanation that made sense was that it must have been intentional.
What followed was a particular kind of suffering. People were bombed by a coalition they could not identify, could not reach, could not hold accountable. There was no one nearby representing the operation, no mechanism for redress, no way to tell their story or have their suffering recognized. In Hawija's case, the bombing did not even achieve its military objective—ISIS was not driven out until two years later. So communities lived in constant terror, bombed by Western powers while simultaneously living under an extremely violent regime. The harm rippled outward in ways that initial casualty counts never captured: lost income, high medical costs, fractured communities, displacement. The impact did not end when the bombing stopped.
Now artificial intelligence has entered this system. The analysis of surveillance data is increasingly outsourced to algorithms and machine learning systems. In Gaza, the Israeli military used a system that marked approximately 37,000 people as having high-risk profiles. Operators had just 20 seconds to verify each designation. The shift was toward quantity over quality. AI turns people into data points, but it also dehumanizes the people making decisions. They become rubber stamps. Sometimes they only check whether the proposed target is male or female. If male, proceed. They are told the system is correct 90 percent of the time—a claim with no verified evidence—but it shapes how they feel about their own responsibility. They blame the machine, even when they can see errors in the data and categorization.
This system has been developed, tested, and normalized step by step: Afghanistan, Ukraine, Gaza, Iran. Within the first two weeks of the Iran campaign, 10,000 targets were engaged. That is unprecedented scale. And the consequences follow well-established patterns. Nine out of ten casualties in urban warfare are civilians. Amazon, Google, Microsoft—they all provide the cloud infrastructure and software that makes this possible. When public outcry emerges, companies pull back temporarily. Microsoft did this with Azure after the IDF's use in Gaza became public. But then another company steps in, or the work continues through different contracts. It never actually ends.
The surveillance does something else too. People know their phone calls and online behavior are being monitored. In Gaza, algorithms were reportedly trained on behaviors like frequently changing your phone or address—things people do in a war zone simply to survive—or being in a WhatsApp group with someone on a watchlist. People don't know what behavior might classify them as high-risk. Gould calls this psychic imprisonment. People freeze. They don't know what is safe to do. Even without drones overhead, surveillance technologies infiltrate everyday communication and life. The harm doesn't stop when the bombing stops. And yet Western governments persist with this approach despite limited political results. Wars, Gould suggests, are not always about winning. Sometimes other things are achieved: political interests, economic interests, access to markets and resources. The military apparatus has become obsessed with perfecting the kill chain—maximizing surveillance data, processing it as fast as possible, engaging as many targets as possible. It has become an end in itself, no longer in service of political objectives. Many commercial companies are feeding that desire, selling an idea of how war is fought and won. To change this conversation requires confronting what war actually is: death and destruction. These are not incidents that went wrong. They are weapon systems built to increase the speed and scale of warfare to unprecedented levels. That is strategy. That is a very different conversation.
Citas Notables
If you have fewer people on the ground, you lose human intelligence. Your decisions are only as good as your data. And if the data is remote, it's probably not that good.— Lauren Gould, associate professor of conflict studies, Utrecht University
We need to confront what war actually is: death and destruction. These aren't incidents that went wrong, these are weapon systems built to increase the speed and scale of warfare to unprecedented levels.— Lauren Gould
La Conversación del Hearth Otra perspectiva de la historia
When you say remote warfare "masks" civilian devastation, what do you mean exactly? The strikes still happen. People still die.
The masking isn't about hiding the strikes themselves—they're documented, sometimes even broadcast. It's about hiding the texture of what those strikes mean. A Western viewer sees a missile launch and an explosion. They don't see the school, the neighborhood, the people who knew the dead. The distance is literal and technological, but it's also psychological. It creates permission.
Permission for what?
To keep doing it. If you never have to see the consequence, if your job is just to verify a data point in 20 seconds, if an algorithm tells you it's 90 percent accurate, then you can keep going. The moral weight disappears into the system.
But surely operators know civilians die. They see the reports.
They do. But there's a difference between knowing something abstractly and understanding it. When I spoke to people in Hawija, they described a kind of double betrayal. They had believed the precision narrative—that Western technology was so advanced it could distinguish combatants from civilians. When that belief was shattered, it wasn't just grief. It was the realization that either the technology failed or the intention was different than claimed. That uncertainty is its own kind of harm.
You mentioned that wars might not be about winning. What do you mean?
Look at what's actually being optimized for. It's not political objectives anymore. It's the perfection of the kill chain itself. How fast can we process data? How many targets can we engage? The weapons exist, so they get deployed. Companies profit from deployment. Investors have financial interest in conflict. The system perpetuates itself.
So how do we stop it?
We have to stop talking about war as a technical problem to be solved and start talking about what it actually is: death and destruction. These aren't mistakes. They're features of the system.