

November 21, 2026
By Yiannis Damellos and Mars Strangelove Jr.
Sources: Reuters, Associated Press
Somewhere in the future, a machine will identify a human being as a target, calculate the probability of a threat, and decide whether that person should continue living. The machine will not hate. It will not hesitate. It will not remember the face. And when someone asks who authorized the decision, the answer may be buried somewhere between a software update, a military procurement contract and a classified operations manual.
That future is no longer confined to science fiction. On November 21, United Nations-backed negotiations in Geneva ended with countries agreeing to continue discussions on regulating lethal autonomous weapons, with the possibility of developing new rules by 2028. They did not agree to begin binding negotiations immediately. The result fell short of the demands of roughly 100 countries seeking legally enforceable restrictions on weapons that can identify and engage targets with little or no human oversight. The United States, Russia and Israel opposed that approach, leaving the question of whether future rules would be legally binding unresolved.
The debate is often described as a discussion about killer robots, but the technology is broader than the phrase suggests. Remotely piloted drones are generally controlled by human operators. Other systems use artificial intelligence to identify objects, track movements or recommend targets while a person retains the authority to fire. Lethal autonomous weapons go further: they can select and engage targets without a human making the final decision in each attack. The central question is not whether a machine can fly. It is whether a machine should be allowed to decide when another human being dies.
The military attraction is obvious. Autonomous systems can operate continuously, process enormous quantities of information, react faster than human operators and function in environments where communications are unreliable. They can also reduce the exposure of soldiers and pilots. In a conflict where seconds can determine the outcome of an engagement, commanders may see automation not as a philosophical choice but as a competitive necessity. If one military deploys systems capable of reacting at machine speed, its rivals will feel pressure to develop comparable capabilities.
That logic is precisely what makes international regulation so difficult. Every government can acknowledge the danger of losing human control while fearing that restrictions will benefit its adversaries. The result is a familiar security dilemma: each side claims it needs the technology because others may possess it, and every new capability becomes an argument for the next one. The Geneva negotiations exposed that contradiction. Governments have not abandoned the idea of rules; they have postponed the moment when rules might restrict what their militaries can build and deploy.
The consequences are already visible in contemporary warfare, where drones, automated surveillance, electronic warfare and algorithmic targeting increasingly shape operations. But it would be misleading to suggest that fully autonomous weapons are already widespread or that every drone attack is conducted by an independent machine. The technology exists along a spectrum, and the degree of human control varies. That is why definitions, accountability and the ability to intervene are central to the negotiations. The International Committee of the Red Cross has warned that autonomous weapons can behave unpredictably and insists that meaningful human control must remain over the use of force.
There is also a problem that no software engineer can solve with a better algorithm: responsibility. If an autonomous system kills civilians after misidentifying a target, who is accountable? The commander who deployed it? The manufacturer that designed it? The programmers who trained its recognition system? The government that approved its use? International humanitarian law already requires parties to distinguish between combatants and civilians and to take precautions to limit harm. But applying those obligations becomes more difficult when the chain of decision-making is distributed across machines, operators and institutions.
The Geneva outcome leaves the world with two years of further discussions while military development continues. That does not mean every country is secretly preparing to hand its armies over to machines. It does mean that the states with the greatest military capabilities have so far declined to accept a firm timetable for binding limits. The negotiations remain open, but so does the path toward increasingly autonomous systems.
Philip K. Dick imagined worlds in which the machinery of power became so pervasive that people could no longer tell where human decisions ended and artificial systems began. The unsettling part of the present debate is that we do not need to borrow his imagined future to see the question taking shape. We are already discussing weapons that may eventually make decisions once reserved for human beings, while governments negotiate over how much control they are willing to surrender.
The machines are not waiting for consciousness. They are waiting for authorization. And the people negotiating their future rules have not yet decided whether they are prepared to deny it.

No comments:
Post a Comment