Here's a somewhat troubling report from NKH regarding how the Israelis are using AI in war.

One red line in AI warfare I've come across is that we're in trouble when there is not a human making the final decision on when to kill another human. This report quite explicitly states that that red line has already been crossed. A human is still pushing the button (or not pushing the ABORT button I imagine), but according to one of the people interviewed, those using the AI have no time to analyze the data the AI has digested to determine if the action they (actually the AI) are taking is appropriate, so they just follow the AI findings without question.

So they're doing what the AI tells them to do. So the AI is making the decisions on when someone should die, and all the people within a certain radius.