STM
Kargu autonomous attack drone from Turkish company STM
“We tried it.” Two years ago, a test was carried out with fully autonomous drones programmed to destroy everything they found in a given area. According to a top figure in the Ukrainian defense industry, there are confirmed casualties.
Fully autonomous drones, without human supervision, killed soldiers on the battlefield for the first time. The revelation, made by a top figure in the Ukrainian defense industry, marks a decisive moment in the history of war.
The test involved 10 drones “Terminator” controlled by artificial intelligence, which were used on the front lines of the war in Ukraine. Two Russian soldiers were killed.
“Let’s experiment”, tells New Scientist Alexander Kokhanovskyya drone manufacturer that provided the technology, at a press conference organized by the Ukrainian embassy. “It was a test. We never implemented it on a larger scale.”
The test took place two years ago and involved drones quadricópteros scheduled for fly towards the front line, travel between 3 and 5 km over the course of about ten minutes and then activate “Terminator mode”in which an AI model searches for and intercepts targets.
“We simply launched it and We know everything will be dead“, disse Kokhanovskyy “Everything found in that specific area will be dead. There is no connection to the drone, you can’t watch video, nothing… Everything it sees will be killed.”
Since there was no way of knowing what the automated drones had seen or targeted, they were sent human-piloted drones to the zone after the test, to manually check the results.
The victims included “a couple of soldiers, a truck”said Kokhanovskyy. Although there is no recording of automated drones attacking these targets, it was concluded that they had killed them.
Kokhanovskyy said he was not personally present at the test, but said it was carried out by a unidentified military unit near the cities of Bakhmut and Chasiv Yar, in the context of a Ukrainian offensive.
The Ministry of Defense of Ukraine did not answer questions about testing nor about the current legal position regarding the use of fully autonomous weapons.
O AI use is common in the military around the world, helping to select targets from huge volumes of information data and automate certain functions of weapons systems. But, at some point in the process, there are always humans involved.
Kokhanovskyy’s admission is apmost categorical proof so far that a death in combat occurred exclusively at the hands of AI.
O Ukrainian government currently bans the use of AI in the final phase of interception of targets, according to sources from defense companies who participated in the press conference at the embassy, although the AI is used in many parts of the process, across multiple devices, up to this point.
Kokhanovskyy affirmed que the Government is aware of AI’s growing capabilities and is holding talks with defense companies about whether or not to relax the rules.
Stories from 2023 suggested that Ukrainian attack drones equipped with artificial intelligence were locating and attacking targets without human assistance — but they would be used against vehicleslike tanks, and not against infantry. At the time, no human casualties were confirmed.
The UN has warned of the risk of these weapons violating international humanitarian law and human rights, by remove human judgment from war. There is also the risk of autonomous systems making errorseither attacking soldiers or equipment from their own side, or hitting civilians.
Most armed forces are developing devices that automate at least part of the target attack process. The US has software capable of gathering and analyzing huge volumes of very diverse data and select targets on the battlefield, which can then be targeted by drones. In theory, however, these processes requires human confirmation.
It has also been alleged that the United States is developing so-called Goalkeeper flying drones and Whiplash naval dronesable to find the your own targets and eliminate them.
A 2021 UN report even suggested that a , produced by Turkish defense company STM, could have been used the previous year to autonomously attack humans.
Mariarosaria Taddeofrom the University of Oxford, argues that killing with AI robs the soldier of dignity, removes responsibility from the attacker and must be banned.
“It’s not just problematic, it’s horrendous“, he said. “Do we want to be the society that kills other people, that allows its Government to kill other people, without human involvement?”
Second Anthony Kingprofessor at the University of Exeter, in the United Kingdom, although fully autonomous attacks without human intervention are technologically possible, they could be a less decisive tool than many think.
“It is certainly possible that governments would allow this if it gave them any military advantage. However, very few, if anyof the millions of drones used in the war in Ukraine by Russian and Ukrainian forces were fully autonomous”, notes King.
“So it’s not just about whether it’s ethically correct to keep humans in the process; at this time, it is also militarily more effective”, adds the British professor.
