In the fourth year of Russia’s full-scale invasion, Ukraine’s war has become a crucible for the technologies of the 21st century. Among the most transformative—yet controversial—tools now emerging on the front lines are battlefield AI systems, particularly those designed to autonomously detect, identify, track, and, in some cases, kill enemy targets. Often dubbed “hunter-killer algorithms,” these systems are rapidly reshaping the nature of combat and raising profound ethical questions about the
Sadly, I’m not surprised.
Both sides have been clearly working on enhanced autonomy for a while now.
At first, it seemed that autonomous targeting would soon remain the only option in face of electronic warfare taking down a majority of drones. (The spectacular footage we’ve seen so far has mostly originated from a small minority of drones that got through. This is changing with fiber optics, of course.)
Then, tactical tricks (flying repeaters) and new guidance methods (fiber optic wire) gave direct guidance a fighting chance again, and somewhat postponed the need for high autonomy…
…but soon enough, an average drone will be capable of much more processing than a super expensive cruise missile from the 1990-ties, and this kind of weapons can be highly autonomous. You can give them the approximate location of a target and tell them to look for something - a ship, a train, an aircraft, a bridge, and of course vehicles with protruding pipes.
It will get nasty and complicated when they get cheap enough to target individual humans, because both common sense and international law insist that humans may be non-combatants and even combatants can surrender. A drone with enough mind to understand will be required to understand this, but there will be a motivation to cut corners. :(