I really don’t get all the fretting about drones making kill-no-kill decisions. You’re telling me that you’d rather this was done by a stressed, emotional, tired front-line private? Reminds me of the debate about a self driving car that might need to mount the curb to avoid hitting a car - and therefore endangering a pedestrian. It’s not an easy decision but I’d rather a machine made it than a stressed person!
The issue is that software making these decisions can't possibly have the full context when making the decision because it was coded long before the decision point occurred. What does that mean in practice? Have you ever saw a flash in the corner of your eye, looked for a second, then realised it was just a reflection? For that split second, your brain identified a potential threat and was trying to quickly decide wh…
.. or vice-versa. The average war has plenty of stressed people firing semi blindly at half-identified shapes. And if we want to be literal, there's the case of Lee Clegg and the exact circumstances in which it is legal or not to open fire on a passing car and kill a teenage girl.