For Russia, the question of robots taking over the role of soldiers on the battlefield is a matter of when, not if.
“Living fighters will gradually begin to be replaced by their robotic ‘brothers’ who can act faster, more accurately and more selectively than people,” Vitaly Davydov told RIA Novosti on April 21. Davydov is the deputy director of Russia’s Advanced Research Foundation, its DARPA analog.
Developing robots for war is an international phenomenon, and Davydov’s remarks fall into a broader pattern of Russia seeking to augment its existing military with new, autonomous capabilities.
“There is a still-classified Russia’s military robotics road-map that sketches out various stages of Russian unnamed military developments, which undoubtedly has been influenced by Russian military actions and experience in Syria,” says Samuel Bendett, an adviser at the Center for Naval Analyses.
While some Russian robots have underperformed expectations in combat, the Ministry of Defence is working on a new generation of combat machines for training and possible future use. At the center of this design is the Marker UGV, or uncrewed ground vehicle. Resembling a miniature tank with treads and turrets, the Marker is as much a test bed as it is a machine expected to see battle.
“The Ministry of Defence is discussing the eventual use of robotic swarms in combat— and Marker is definitely the platform to test that out,” says Bendett, an Adjunct Senior Fellow at CNAS. “As envisioned, it will be able to launch swarms of UAVs or loitering munitions, making it a truly versatile robotic platform.”
The development parallels many robotic programs under way in the United States. It is remarkable, too, that both nations have hit upon swarms of ground robots as a way to supplement existing human formations in combat. Russia plans to test swarms of ground robots later in 2020.
Part of Davydov’s argument for military robots is that they will, simultaneously, be faster, more discriminating in target selection, and more accurate than people. While it is already possible for machines to react to sensor data faster than humans, the other parts of that equation require targeting algorithms with far greater discretion than anything seen presently. (It also involves sidestepping the entire international legal and moral debate over transferring the decision to kill in battle from humans to machines.)
“At this point in military autonomy, a human can designate targets once they are identified,” says Bendett, “but once the computing speed and the levels of battlefield complexity multiply exponentially, I wonder how true Davydov’s statement will be.”
Making robotic soldiers that adhere to the laws of war and protections for civilians isn’t just an ethical challenge, it’s a coding challenge. It requires the military to prioritize positive target identification over the fastest possible response.
“At this point, the Ministry of Defence says that robots replacing humans saves human soldiers from danger,” says Bendett. “Built into that statement is the assumption that a sophisticated unmanned system would be able to eventually distinguish military targets from civilian ones, avoiding unnecessary casualties.”
It doesn’t help much, either, than the militaries responsible for pursuing this kind of battlefield autonomy already have atrocious records when it comes to avoiding civilian deaths.