I always have always believed that the future soldier is not a human. Fly a few of these inside North Korea and let them take out some high ranking generals. Then, drop a note simply stating "GOTCHA, YOUR NEXT, FAT BOY. DON'T GO TO SLEEP"! We'll let our robots fight our ground wars, air wars, under sea wars and space wars.
As a veteran as well as a long time (Since 1993) user of the internet, having seen up close and personal the meaning behind the phrase: "This is a military operation. Nothing ever goes according to plan....", along with "Murphy's Law": "Anything that can go wrong, will go wrong at the worst possible moment....", I am leery about not only arming drones, but giving them "autonomy" to decide when, where, and how, to engage "the enemy". We have yet to develope computer software sufficiently secure enough to prevent malicious hackers from compromising our online banking network, retail sales, utility infrastructure controls, and even routine communications. On top of that, we haven't even proofed the software from routine corruption from normal usage (Anybody still using Windows 3.11???). Hell, after seventy-plus years of use, we still haven't figured out an effective method to synchronize traffic signals on our city streets. And we're supposed to turn loose a drone armed with a sniper rifle, grenade launcher, maybe at some time in the future a high-powered laser, and let "it" decide who to "eliminate" along with when, where, how, and why???!!! Get back to us on this AFTER we develope a working "morals", and "conscience" subroutine for the "artificial intelligence" program....