This article originally appeared on www.bibliotecapleyades.net in Spanish. I have converted them to English
by Jeronimo Rajchenberg and Joaquin Salam
April 17, 2026
from the Alai Website
Remote killing and destroying instruments are nothing new:
the idea of inflicting harm without exposing oneself to danger has been around since the time when a stone was thrown or a crossbow was used.
Over time, remote devices have been perfected and have become more efficient, turning the crossbow into a remotely controlled drone and creating a link between technological advances and military devices.
As worrying as it may be, the story doesn’t end with drones.
Over the past few decades of rapid technological development, an effort has been made to familiarize people with the idea of highly technological weapons in numerous popular media, from Hollywood to science fiction writers and other actors who transform culture, so that the public (that is, the “world population” – as they see it – or the potential market) is more likely to accept the new reality when it becomes tangible.
The public imagination has been fueled in two different directions:
On the one hand, a narrative to convince people that technology will bring about a new golden age for humanity, that with the incorporation of new technologies, all worldly work will be done by robots, and humans will be free to become philosophers, musicians, painters, etc.
Following this narrative, technology will solve problems of resource limitations, freeing the world from wars over its control.
Star Trek could be an example of that way to present this new utopia, in the style of the “end of history”, of a future without scarcity.
On the other hand, there is the narrative of the dystopian future in which practically indestructible robots provide a very small group of people, or the robots themselves, with the ability to kill or enslave humans on a large scale.
The best-known popular examples include Robocop, Terminator, and Matrix.
Unfortunately, there is much more evidence and evidence that the world is moving towards this last option. In practice, it seems that automation, and with it advances in technologies similar to AI, seem destined to reduce that promised freedom.
The increasing abandonment of human labor in favor of automated substitutes is leading to an increasing mass of the population being considered potentially “expendable” by the oligarchs who govern it.
This is not freedom from work, but rather it increases the number of people living in poverty while maintaining the current economic model.
In 2025 alone, in the United States, around one million jobs (i.e. people) were replaced by A.I.
In other words,
Advances in A.I. seem to be based on the deepening of a model of privileges that allows a remarkably small portion of the human race to dispossess the rest.
A Lethal Combination
The modern combination of A.I. with robotic development results in what are known as “killer robots” or “automated weapons systems” (AWS).
Killer robots possess capabilities that a human soldier could never develop, such as greater resistance to kinetic attacks and the absence of fatigue; the absence of stress, panic or any form of empathy, allowing for an uninterrupted flow of commands or “orders”.
Added to this is the acquisition of “in the field” data.
Most machines already deployed have the ability to evaluate a large amount of information, especially when operating in swarm attacks, where thousands of cameras and sensors generate a myriad of visual perspectives and evaluate a large number of decisions that an individual or even a group of humans could never evaluate or process.
The adoption of the swarm attack protocol is not only a change in the mass of assets, but a change in the conception of war.
Swarm protocols are not just a massive expansion of the number of soldiers in a battalion, so to speak, but a completely different way of conceiving, commanding and executing a particular task.
As such, they require and are possible thanks to a gigantic increase (beyond human capabilities) in data collection and the processing of information and intelligence.
The idea of a ‘killer robot’ can be considered a totality of the logistics infrastructure, from the storage of data and its processing to the final execution of orders by an autonomous machine, which provides new data for the cycle to continue.
The ethical, humanitarian and legal dimensions of killer robots,
become truly problematic when humans provide machines with lethal autonomy, that is, the ability to make decisions about what constitutes a military objective and to act accordingly.
When machines are not tools operated by humans, such as the crossbow or even the remote-controlled drone, but rather take ownership of the decision-making process, they become the brain and humans become the workforce.
Two non-lethal examples of this are,
the AI-oriented process that Amazon has put in place by allowing it to make decisions about hiring and firing people
the rentahuman.ai Web site, where A.I. agents hire people for specific tasks
In both examples, the narrative that machines will work for humans is dismantled.
The Present
Although science fiction at the end of the 20th century and the beginning of the 21st leads us to imagine killer robots as humanoids or quadrupeds similar to dogs (versions that, in fact, exist), this is not the only form that killer robots can take.
Below are some examples of real killer robots that are already in operation and that don’t fit this description.
As initially reported by Yuval Abraham on 972mag.com, the Israeli system based on military A.I., Lavender, was used to designate people suspected of being members of the military wing of Hamas and Palestinian Islamic Jihad as potential bombing targets, with about 20 seconds per target for human “supervision” – which consisted mainly of verifying only that the target was a human being.
This was done knowing that “mistakes” were made in approximately 10% of the cases.
These targets were then bombed not on the battlefield but, in most cases, in their homes with their families, causing surrounding civilians and entire families to be designated as “collateral damage”.
Where’s Daddy is a similar system – its name reflects the level of humanity behind its design – used to identify when an alleged “militant” entered their family home, and then ordered the attack, which practically guaranteed civilian deaths, with the total destruction of apartment buildings.
The Gospel is another A.I. system used by Israel to identify suspicious buildings: described by a former intelligence officer, as a system that essentially facilitates a “mass murder factory”.
In the first six weeks of the Gaza genocide, the Israeli army killed more than 15,000 Palestinians, relying heavily on A.I. for target selection.
The fact that it’s a militarily programmed human being who pushes the button at the end of the chain doesn’t make the whole process look less like a killer robot.
Red Wolf is an experimental facial recognition program, deployed at checkpoints in the Israeli-occupied city of Al-Khalil (Hebron), which collects scanned Palestinian faces and adds them to a database, which determines who can cross a checkpoint.
The ultimate goal is to eliminate human intervention to further facilitate the automated apartheid that this imposes.
The Israeli occupying forces have a corresponding application called Blue Wolf, which generates classifications based on the number of registered Palestinians, which implies, in practice, a gamification of the application of the apartheid system.
This goes hand in hand with the development and testing of automated turret systems at a military checkpoint in Al-Khalil, such as Smart Shooter, an autonomous turret capable of shooting at people without a human pointing or pulling the trigger.
The combination of these systems – Red Wolf and Blue Wolf together with Smart Shooter – constitutes an example of a lethal military system that works almost completely independently of any human intervention.
This is one of the first examples of lethal autonomy in which an A.I. system is forced to kill based on its own “criteria”.
Even without questioning the system’s human programming ethics, granting machines the ability to make decisions independently is problematic because of the “mistakes” they can make, also known as false positives.
The Flock AI system, for example, misidentified Brandon Upchurch in Toledo, Ohio (USA), as the driver of a vehicle with stolen license plates.
Mr. Upchurch was then the subject of a violent arrest and an attack by a dog because the system misread a number on his license plate and triggered an alert.
