re: #184 goddamnedfrank

Think about a simple hypothetical. We send an automated XB 47 on a ground attack mission against a specific target, but along the way it is engaged by an air defense radar system. It’s an expensive platform and sooner or later as we invest more and more money into expanding the capabilities of unmanned systems it’s going to be very tempting to give them the ability to react defensively to events in their environment. So let’s say we’ve armed it with anti radiation missiles and given it permission to automatically target hostile ground radar installations that engage it along the way. This might seem reasonable at first blush, except this radar system is located near a school, hospital or other civilian target.

We deal inexpertly enough with collateral damage and consequences when it’s inflicted by human judgement calls. So again, considering the trivial nature of the software updates necessary, the cost of the systems at risk and political fallout potential, will we even be allowed to know when machines are allowed to autonomously decide upon specific use of force actions?

The answer to the question of the bolded portion is in my mind to wave it off. It is the enemy’s fault for placing a military system near civilians like that, and thus the blame for whatever happens is on them.