The Hidden Human Cost of Algorithmic Warfare

Fresh from their “snuff-movie” hit incinerating Venezuelan fishermen, Team Trump moves yet another carrier strike group into the Persian Gulf. Suddenly, our infotainment airwaves are full of experts spruiking “clean, surgical strikes,” while our media eagerly repeats the Pentagon’s propaganda. An old fat sea-cow, the USS Abraham Lincoln, and her tattooed bouncers are framed as instruments of precision and humane restraint, hovering just over the horizon of Iran’s ruggedly spectacular coast.

“Surgical strikes?” Pentagon experts now propose to kill and maim Iranians in an illegal blitzkrieg or perhaps three months of “boots on the ground”—the messages are as garbled as a Trump rally speech. But what is clearly being sold is the old lie that war is glorious, noble, and heroic. The US is supposedly ready to “send a message” without another Iraq-style quagmire because, this time, war will be data-driven, algorithmically optimised, and somehow morally minimised.

Modern warfare has never been more complex, nor more bloodthirsty. Today, to hold a principled anti-war stance is often derided as “un-Australian” or weaponised through accusations of anti-Semitism, all while a new cycle of state-sanctioned Islamophobia plays out under the guise of national security. We are witnessing the return of “One Nation” rhetoric: a toxic mix of division and rabid ignorance. From the White House, the lies arrive with such velocity that they overwhelm the public’s ability to process them. Above all, we are sold an antiseptic fantasy: that the next war will be a clean victory won by Artificial Intelligence, where autonomous drones and “algorithmic warfare” replace the messy reality of human slaughter.

We are rarely told who taught the machines to kill. And at what human cost.

The reality of 2026 is that the “intelligence” in AI remains deeply, painfully, and inexorably human. AI-enabled targeting, surveillance, and logistics systems require billions of data points to be labelled, sorted, and refined before a single model can be deployed. Every box drawn around a body in a blurry image, every classification of rubble, every tag of “weapon” versus “non-combatant” has been performed by a human being. Not by Silicon Valley engineers, but by a vast, hidden army of pieceworkers scattered across the Global South.

In refugee camps in East Africa, in cramped internet cafés in South Asia, and in crowded apartments in Latin America, workers are paid the equivalent of a few dollars an hour to sit at flickering screens and trace rectangles around human silhouettes. Behold the invisible pedagogues of the war machine, providing the labelled examples that allow military AI to distinguish “target” from “background,” “combatant” from “crowd.”

The irony is dark and palpable. Many of these workers live in regions already wrecked by Western interventions. Some fled earlier conflicts in Iraq, Syria, or Afghanistan; others live under permanent austerity. Men and women now find themselves training systems that may one day patrol their own skies. It is a grim circularity: the global poor—the “wretched of the earth,” as Frantz Fanon termed them—are pressed into teaching the next generation of weapons how to see.

This is the new “Digital Taylorism.” Just as 20th-century manufacturers broke down manual labour into minute, repetitive tasks, 21st-century AI firms have fragmented intellectual labour into atomised micro-gestures. For those training military models, the work is often traumatic. Investigations into data-labelling hubs in Kenya, India, and Colombia document the harm: workers are forced to view thousands of hours of violent, graphic content—war footage, torture, and the aftermath of bombings—to “fine-tune” the algorithm’s recognition.

Unlike the soldiers who will eventually operate these systems, these digital labourers have no veteran status, no medals, and no guaranteed access to mental health care. When their performance drops due to the trauma, the solution is simple: deactivate their account and hire another worker from the endless queue.

Australia is not a bystander. Firms like Palantir and Anduril have successfully blurred the lines between civilian and military data. In February 2026, the Labor government quietly awarded Palantir a fresh $7.6 million contract for Defence’s Cyber Warfare Division. Meanwhile, Canberra has committed $1.7 billion to Anduril’s “Ghost Shark” program—autonomous undersea vehicles designed for strike operations.

When these systems are woven into civilian infrastructure, the war machine becomes an everyday reality. The same optimisation logic used to squeeze more deliveries out of a warehouse worker is repurposed to accelerate the “sensor-to-shooter” loop. In Australia, we saw a prototype of this in Robodebt: the weaponisation of data against the poorest, treating them as problems to be hunted by algorithms long before any human looks at the facts.

This is not a glitch. It is how capital has integrated AI into the security state. A data labeller in Nairobi might make less in a day than a single second of flight time for a carrier-based fighter jet. The system depends on the invisibility of the connection between the micro-task on a screen and the missile in the sky.

We must refuse the comforting illusion that the coming war will be “clean” because it is “smart.” If our automated future is built on a foundation of traumatised, underpaid labour, then it is not a technological triumph. It is a moral failure disguised as innovation. The cost of the next war will not only be counted in missiles fired and lives lost in Tehran or the Strait of Hormuz. It is already being paid, quietly, in the human dignity we have sacrificed to train the machines that will fight it.

Coda: The Sycophant’s Algorithm

And so, we find ourselves back in the familiar, fawning posture of the Australian security establishment—a collection of strategic wallflowers so desperate for an invitation to the dance that they have handed the keys to the kingdom to a band of Silicon Valley carpetbaggers. We are told that by tethering our national interest to the likes of Palantir and Anduril, we are buying “security.”

In reality, we are buying a front-row seat to our own irrelevance.

We have become the regional branch managers for a war machine we neither control nor understand. To watch a Labor government—the party that once spoke of “national sovereignty”—quietly outsource our military intelligence to foreign algorithms trained by the global dispossessed is more than a policy failure; it is a spiritual surrender. It is the triumph of the technocrat over the citizen, the dashboard over the diplomat. We are being marched into a conflict in the Middle East not by the force of reason, but by the relentless, unthinking click of a mouse in a Nairobi sweatshop. It is a spectacle of profound hollowness, orchestrated by people who wouldn’t know a national interest if it bit them on the leg in the middle of a Canberra cocktail party.