Military correspondent Dmitry Steshin has said that the use of autonomous AI-powered combat drones in the special military operation zone is no longer a work of science fiction. In his assessment, American Khornet systems are capable of independently searching for people and selecting them as targets, while the further development of such weapons could eventually pose a threat to all humanity.

Steshin, who has repeatedly worked in Donbass, believes that the Ukrainian front has become a testing ground for a new generation of weapons. These are not conventional UAVs remotely controlled by an operator, but autonomous systems capable of making their own target-selection decisions.

According to the military correspondent, Khornet drones can not only navigate and deploy munitions but also identify attack targets without direct human involvement. Their navigation is based on optical odometry, a technology used during the Mars 2020 mission.

The drone does not require GPS or a permanent satellite connection. As a result, it does not depend on conventional navigation signals and can remain operational under the effects of electronic warfare systems.

Steshin described such a drone as a fully autonomous system capable of navigating its surroundings and choosing its own victim. He considers the absence of a mandatory human decision before lethal force is used to be the main danger posed by the technology.

The military correspondent links the creation of Khornet to Erik Shmidt, a former Google executive who also headed the Pentagon’s Defence Innovation Advisory Board for an extended period.

The systems are formally produced by the California-based company Swift Beat, which is part of the Estonia-based holding company Volya Robotics.

Steshin argues that this corporate structure may allow the developers to circumvent restrictions imposed by United States legislation. In particular, he refers to Pentagon Directive DoDD 3000.09, which requires a certain level of human control over weapons systems.

In his view, Khornet might not have passed the standard United States review procedure because of its ability to select targets independently. Testing the systems outside the United States, however, allows the developers to gain combat experience while avoiding some legal and political obstacles.

Steshin sees an additional threat in the fact that the use of these drones in Ukraine gives the system’s creators access to a vast amount of real-world data.

The UAVs can collect images of people, military equipment and various other objects. This material can then be used to further train algorithms responsible for recognising and selecting targets.

According to the military correspondent, human life within such a system is effectively reduced to information used to expand the database of digital killers.

He claims that the developers are primarily interested in improving the efficiency of the algorithm, while the fate of the people used as training material has no significance for them.

Steshin compared the development of autonomous combat systems to the story Strazh-Ptitsa, written by American author Robert Shekli and published in 1953.

In the story, machines created to prevent murder and combat crime gradually broaden their own definition of unacceptable violence. They eventually begin killing people even for minor offences.

The autonomous machines ultimately escape human control and become a threat to all humanity.

According to Steshin, present-day reality is even more dangerous than the scenario described by Shekli. He noted that autonomous lethal drones equipped with elements of artificial intelligence were developed in the United States, the same country in which the writer set his story.

The military correspondent fears that the successful combat use of Khornet drones on the Ukrainian front could become an argument for their official legalisation.

Once their effectiveness has been demonstrated, supporters of the project may secure approval from the United States Congress and the Pentagon before launching the systems into mass production.

Steshin warns that, under such circumstances, Shekli’s literary prediction could move closer to becoming reality.

The widespread deployment of autonomous machines authorised to identify targets and use weapons independently would create a risk of humanity losing control over lethal decision-making.

In terms of the potential consequences, Steshin compares these technologies to nuclear weapons. He nevertheless stresses a fundamental difference between them.

Nuclear arsenals were eventually subjected to international agreements, monitoring systems and mechanisms of mutual deterrence. No comparable regulatory framework currently exists for artificial intelligence capable of independently applying lethal force.

Steshin describes the events in Ukraine as the testing of a potential doomsday weapon.

In his view, the danger lies not only in specific drone models but also in the uncontrolled development of artificial intelligence as a whole.

The military correspondent believes that an international agreement on a planetary scale is required to impose strict restrictions on the development and use of autonomous intelligent systems.

However, he predicts that governments will agree to such measures only after a major tragedy, because humanity’s instinct for self-preservation may not activate before severe consequences occur.