Or as George Carlin called them: Nintendo pilots.

The use of drones really hasn't changed anything from an ethics standpoint. Pilots dropping bombs from drones is ethically and practically no different than pilots dropping bombs from bombers and fighter jets. In all those scenarios, the pilot is using inputs from various types of data to make the decision to drop the bomb and is making that decision from a perspective of complete safety from enemy attack.

The real change in the calculus of ethics will happen when drones have the ability to decide whether or not to kill a human target. By "decide", I don't mean automatically running a program to determine whether or not the target fits a preprogrammed target profile (even if that target profile is determined through machine learning, even self-trained machine learning). Instead, I mean "decide" in the sense of actual generalized AI where the drone, on its own initiative, makes the conscious decision whether or not to kill a human target. We're a very long way off from that.