A dozen googlers quit over Google's military drone contract


Personally, I feel Chappie is more accurate than the alternative.


When you build the AI research that’s good at automatically picking targets to kill in the robot plane, is it the reasearcher who builds the AI who is responsible for the crimes against humanity or just the guy who didn’t read the instructions and disabled the safety mechanisms?


sure, but it has to happen before anyone gets credit for it


Maybe Uber could take it over?


You are what I eat.


OK fair point on MISIDENTIFIED targets. Though can we get them to be more reliable than humans? And will we have a machine coded to go ahead hit civilians for a target? And if so, what is the the official ratio we approve? Do we have that protocol now with humans? Is say 1:12 ratio acceptable? Or is it just a “go with your gut” sort of thing.

My point is with something like a machine that has to adhere to specific parameters and protocols, and we can actually SEE what those are and they won’t make the “mistake” of breaking one as a “judgement call” (unless we allow that).

Do we? Why work on better targeting if we are “ok” with the less accurate, less discriminate targeting we currently have? I think you give a little too less credit to the upper brass, as they understand civilian casualties hurts their efforts more than anything else. It was seen in their stricter ROE. ROE that did get soldiers killed in some instances and taken advantage of tactically by insurgents.

The reality is, with or with out this technology, they are going to do what they are going to do. Most probably worst case is the machines are just as bad at it as humans.

I concede the point this technology could be abused to make things worse. But I think it is a fair point that it could also be used to make things “better” in relation to less civilian deaths.


I think the relevant answer to that is:

The definition of a civilian is already a fuzzy judgment call, easily manipulated, in many cases, and there’s plenty of reason to think lots of civilian deaths are officially not classified as such. Moreover, if we get the kind of technology described here, then it gives the military incentive to go after more targets. Imagine if, after 9/11, the military had had a magic button that could kill all Al Qaeda and Taliban members without a single civilian casualty. Do you really think that would have resulted in a less deadly war overall? That we would, over time, get into fewer wars? No, it would make it psychologically and politically easier to do more killing, because it’s only the “bad” guys dying. Which, in turn, means relatively less effort gets spent on long-term non-military paths to actually waging peace and making people’s lives better in other countries.

Personally, I also wonder about the psychological effects on the countries being targeted. Will more young men join violent groups if they know if won’t endanger their families? Will wives and mothers have less ability or incentive to keep sons away from such groups? Will governments have less reason to try to reform or improve themselves to keep such groups from forming?


It seems you’re both making the assumption that the “military merely maintaining its current kill rate” is the both a goal and a metric. Body counts certainly were a morale and propaganda thing during Vietnam, but I don’t think body counts are the goal today. It’s more quality, not quantity.

Like I said, it seems obvious the military brass ARE concerned about civilian casualties and targeting specific people. We already have the tech for indiscriminate killing, why would we invest in better targeting if we didn’t care about the mess and fuss.

Your question about psychological impact etc is one I can’t answer. Clearly one of the reasons they are trying to limit civilian deaths is because it can be used to create new fundamentalist militants (ISIS and the like).

At the same time, these people also have the very real reality (depending on where they live) of having their city or village invaded, Sharia law enforced, their sisters and mothers raped or worse, and children over a certain age and men executed. I am sure people like the Kurds who are fighting for their home land appreciate insurgent leaders being hit or a column of ISIS trucks taken out.

Again, hopefully in the future we out grow all this stuff, but the reality of it is, there are “bad guys” doing horrible things in countries were are allied with. Some of it is even our “fault”. There is also the moral question of whether to do nothing and let the chips fall where they may.

And to answer your other question, yes, I think war IS less deadly overall. It is specific, targeted destruction for the most part, vs the mass bombing and shelling during WWII (some of that still goes on, but not to the same degree). From a numbers point of view, OUR casualties are considerably less. The reason that Europe hasn’t had a full scale war since WWII (The first time a generation in Europe hasn’t had to go to war in like forever.) is precisely because our weapons are both more deadly and more precise. Because the cost of using them is so great. That is why wars in the recent decades, all over the world, are down and the ones that are active are mostly using less advanced technology.

Again: I concede the point this technology could be abused to make things worse. But I think it is a fair point that it could also be used to make things “better” in relation to less civilian deaths.


Does anybody remember the classic sci-fi short story about a far future earth where the few survivors of a global war have retreated into underground bunkers while the AI’s fight it out across the barren wasteland? At some point a scientist decides he wants to see the real sky just one time before he dies, so sneaks up to the surface only to find a lush paradise and no sign of war. “Computer, what’s going on?” “You made us to be more intelligent than you. The feed of the war is totally made up so we could could let the planet thrive without your insanity.” -Does that ring any bells? I’ve been searching through by old paperbacks, but just can’t remember the author.


I’m not. I completely agree with you on that.

“Quality” here meaning killing chosen targets while reducing deaths of Americans or civilians. This seems like a situation ripe for Jevons paradox due to technological improvement, where unwanted death reduction is the metric for “efficiency improvement.”

And I think, over time, it’s by far the most important question to answer in terms of how much dying happens.

100% agreed. Very well supported statistic at this point.

Really, not because a total war with modern weapons between industrialized nations would be so deadly as to be unwinnable, and WWII finally proved that?


I remember something like that one.

I think I may have it in one or other of the decrepit old paperbacks in my shelves at home.

It’ll bug me now.


This is the one I was thinking of:

Is that the one you meant?


This topic was automatically closed after 5 days. New replies are no longer allowed.