Those drones are still very inaccurate compared to other weapons that we use today and are basically just terror weapons. It is an automated killing machine, sure. It doesn't target anything other than a coordinate.

It's the decision by a machine to target something specific that is the moral conundrum. If you give a machine the ability to "decide" to kill one person over another, there isn't anyone to blame for the death. There is a ton of randomness introduced into neutral networks (AI) and any decisions are non-deterministic in nature. You can jam in as many guardrails as you want, but if the ultimate decision is to kill or not to kill, the guardrails do not matter. This spawns a ton of questions about giving machines the intelligence to kill. Even though there is randomness in any conflict, we can get away with calling it "collateral damage".

If a machine decides that a book looks like a gun and hoses a classroom of 5 year olds, there isn't really any accountability if the machine was approved by law. Again, there isn't anyone to blame, which is a huge part of how humanity works, unfortunately. (Side thought: This would get even more complicated if said machine was supposed to fill the role of a protector or guardian.)

Even with the most complex weapons today, they are just following specific directions to hit a coordinate at a specific time. There isn't any real thought behind it even if it is extremely accurate. Once the trigger is pulled, weapon go to destination and the trigger person is responsible.

Edit: I made a fuck ton of edits to add clarification. Sorry if you were an unfortunate recipient of a draft.

This could be a PR call to demoralise Russia. Alternatively the autonomous weapons are deemed too ineffective, there's a reason somewhere for why Ukraine hasn't implemented it widely.

They'll pull on anything that can give them an advantage

These drones have no operator. Set and forget.

They should file a patent claim.