Human Rights Watch wants to ban armed autonomous robots

[Admin]

I'm not sure if you guys have seen this yet, but HRW is trying to ban the Terminator.

I haven't read the report yet because I'm coincidentally still reading this book and want to finish it first:
Governing Lethal Behavior in Autonomous Robots

But after reading the press releases and online commentary, I'm much under the impression that the HRW report was written with little input from robotics engineers.

My personal opinion is that humans are historically unethical on the battlefield. It's hard to name a single conflict that didn't have a massacre of unarmed civilians. If a robot can be programmed to not kill out of hatred (or whatever other reason), I would argue it to be unethical to *not* use a robot soldier instead of a human soldier.

[sys 49152]

Whatever happened to Asimov's 3 laws? 

[Admin]

Whatever happened to Asimov's 3 laws? 

Asimov's stories showed how his laws didn't work

[Soeren]

Hi,

If a robot can be programmed to not kill out of hatred (or whatever other reason), I would argue it to be unethical to *not* use a robot soldier instead of a human soldier.

The problem is not the lack of emotions, but rather the situations where they go haywire and we all know that a human brain (even a not so bright one) is far superior to what a robot is, and will be for a lot of time, when it comes to eg. "pattern recognition" and it's easy to imagine a scenario where even a light damage (bullet, frags or whatever) makes a robot go haywire and then civilians will be in real danger.
And what's the point if two sides each sends robots and no humans (in some future battles) - It reminds me of the situation where two kids argues over who's father can beat the other ones father

I know a lot of you yanks are spoonfed from birth with this "patriotism" or what should I call it and everyone goes on about how the US military defends your way of life and so (apparently a lot of this defending have to take place in oil producing countries ) and if something, this sort of "mass thinking" is something to fear, as that means a lack of questioning the purpose and presents a hook to hang your personal ethics, when enrolling the "common ethics" without a grain of doubt.
Sure I know that I'll be called a lot of things for this, mainly from those who bought into the "come right in and leave your ethics at the wardrobe girl, we have one size that fits all"-crowd, but that's very often the price of free thinking, so I'll live with that

The question of killer-bots or -nots really can be simplified...
How would you (you, the reader) feel about having an armed and powered killer 'bot in your daily life? Would you trust it to never make a mistake and be strangely error free, contrary to all other electronics devices you've ever used?
Even the fairly harmless (in comparison) Robot Wars 'bots have, by rules, a hefty kill switch/link that is only to be removed when entering the arena. A fullblown and powered combat 'bot (combot?) with a couple of miniguns, rockets and what not woud scare the living daylight out of me - even if I had done all the electronics myself - Murphys Laws applies everywhere!

Anybody that (honestly) wouldn't mind sharing their life with such a machine have my full blessings to advocate them (plus my sympathy towards their stupidity )
All others should concentrate on building 'bots that are not ment to intentionally kill, maim and curse.

Funny thing is, people advocating such always start their dogmatic outburst with "I really don't like war, it's terrible, but...".
If all engineers simply didn't enter into developing killer 'bots, guns, bombs etc. there wouldn't be any such devices.

Needless to say, I understand the HRW's worries and I fully support their ideas!

[Admin]

The problem is not the lack of emotions, but rather the situations where they go haywire and we all know that a human brain (even a not so bright one) is far superior to what a robot is, and will be for a lot of time, when it comes to eg. "pattern recognition"

As long as this is true, robots will never be fielded in those situations. The above article is for the hypothetical day when robot brains are actually superior.

and it's easy to imagine a scenario where even a light damage (bullet, frags or whatever) makes a robot go haywire and then civilians will be in real danger.

Humans don't even need light damage to go haywire. My Lai, for example.

And what's the point if two sides each sends robots and no humans (in some future battles) - It reminds me of the situation where two kids argues over who's father can beat the other ones father

When wars are fought over resources (such as oil fields or territory), the victor keeps it. When it is fought over ideals, the victor gets to force their ideals on the loser. etc. etc. etc.

And history shows the better equipped force is much more likely to win.

Funny thing is, people advocating such always start their dogmatic outburst with "I really don't like war, it's terrible, but...".
If all engineers simply didn't enter into developing killer 'bots, guns, bombs etc. there wouldn't be any such devices.

Did you see the monetary prize for the latest DARPA robot competition? And there will always be engineers who believe what he is doing is right because it is for his nation.

I'm not an advocate of war, or making weapons. I refuse to make or assist in the creation of killer robots (yes, peaceful robot tech can be used for evil). But I realize that wars are not going away any time soon,  that dictators won't step down if you say 'please', and people will wage those wars regardless of the weapons they have. HRW can't ban robots no more than they can ban guns or war itself. It's a political - not engineering - issue.

Yes, chemical weapons were banned, and same with cluster munitions, but only because they couldn't discriminate between combatants and civilians. Carpet bombing was ended with the development of the smart bomb. And my argument is that future robots will be better than the current option.

Besides, how are we going to fight off the alien invasion if we don't have giant fighting robots?

[Soeren]

Hi,

The problem is not the lack of emotions, but rather the situations where they go haywire and we all know that a human brain (even a not so bright one) is far superior to what a robot is, and will be for a lot of time, when it comes to eg. "pattern recognition"

As long as this is true, robots will never be fielded in those situations. The above article is for the hypothetical day when robot brains are actually superior.

And as long as your response is true (which unfortunately, I don't believe), we don't have to worry over killer-'bots at all

Humans don't even need light damage to go haywire. My Lai, for example.

Well, we don't have to go back to the Summer of Love... Just look at Quantanamo bay!
But... Such people is not lightly damaged - they're seriously deranged!

Any such acts of sinle persons would net them the instant predicate of sociopath.

When they're part of an army, they're judged differently and while I can somehow understand the mechanisms that makes shell shocked (What they call PTSD these days) people act outside the borders of decency, I'm sure it would be a lot cheaper to develop screening methods to keep such people out of the army, than developing new weaponry, whether mechatronics or not - if the will was in place.

When wars are fought over resources (such as oil fields or territory), the victor keeps it. When it is fought over ideals, the victor gets to force their ideals on the loser. etc. etc. etc.

Wars are always a mix of the two to some extent, but does that mean that you consider Desert Storm to be a U.S. fiasco? I don't see Stars & Stripes in Iraq, so they won?
(And now you're probably gonna tell me that it wasn't about oil - You have very brilliant spin doctors, so most US citizens do believe it to be an act of "liberating" another country - No, we won't mention Tibet and their blinding lack of resources, or the fact that if they had oil, they would have gotten the help to kick out the chinese

Whenever there's a crisis, start a war far away to take the citizens mind off the problems at home. Maggie started the Falkland War, and I guess that has been part of what drove some of the US campaigns as well.

And if you just want to spread western ideology, The World Bank is the major instrument used.

Without the US colonialism behaviour towards oil producing countries, I think you'd have been spared the 9/11 (and that's assuming that none of the conspiracy theories about the incident are valid).

And history shows the better equipped force is much more likely to win.

Well, the French was better equipped than the Vietnamese and so was both the Chinese and the US, but despite that (and the occasional massacres), it didn't happen.
The Russians didn't have much luck in Afganistan back in time either, so don't expect a war to be won by gadget superiority - And the more complexs machinery gets, the more can go wrong.

Did you see the monetary prize for the latest DARPA robot competition? And there will always be engineers who believe what he is doing is right because it is for his nation.

"I really don't like war, it's terrible, but..."
- I wanna get rich
- I let my government tell me what I have to think (even when I know it's wrong ('cause I like money more than decency).
- People with ideologies diverging from that of my countrys/mine are wrong.

Engineers without the common decency should be forced on a three month tour among war victims in the hot spots. I'm sure that would at least bring down their numbers.

I'm not an advocate of war, or making weapons.

Yeah, that's the standard opening sequence

I refuse to make or assist in the creation of killer robots (yes, peaceful robot tech can be used for evil).

Yes and that's the reason I have been silent on the DARPA - Sure, at first it's a rescue 'bot, but when they have something capable, the jackhammer will be exchanged for a minigun I'm sure.

But I realize that wars are not going away any time soon,  that dictators won't step down if you say 'please', and people will wage those wars regardless of the weapons they have.

That has been the MO so far, but it doesn't take a rocket scientist to see that casualties have gone up with the input from engineers - Russian exploding toys come to mind, Napalm, Agent Orange, cluster mines etc. etc.
It's no big secret that war is about casualties - bog down the healthy with wounded that needs care and resources - but as long as nobody even try, it ain't gonna change.
Civilisations that can send a robot to Mars should have the capability for decency as well.

It has been said that a society should be judged from how they treat their poorest subject, but I'd rather judge them on how they treat their enemies - that's what show the amount of decency of the rulers.

HRW can't ban robots no more than they can ban guns or war itself. It's a political - not engineering - issue.

And here you just used the hook I mentioned!
Engineers are people.
Politics is opinions of people
Ergo, there's no separating the two and saying it's politic is a lame escuse, if you pardon my French, to hang your personal ethics at afforementioned hook to join the madness of weapon development.

Yes, chemical weapons were banned, and same with cluster munitions, but only because they couldn't discriminate between combatants and civilians.

There's still a lot of weapon that can't discriminate.

Carpet bombing was ended with the development of the smart bomb. And my argument is that future robots will be better than the current option.

Carpet bombing  was ended because smart bombs is a cheaper way to reach a goal, with less protests from people against war.


And the killer 'bot... How's it going to discriminate between a civilian and a soldier?
People wearing a green jacket and a hole punch may have been forced to do so and guerillas may look like civilians - how should it define who's who, when it's  sometimes impossible for humans?
I imagine a farmer with a rake or similar may easily be mistaken for a guerilla with an MG.

Besides, how are we going to fight off the alien invasion if we don't have giant fighting robots?

Going back to your postulate about superiority:

And history shows the better equipped force is much more likely to win.

I'd imagine that aliens capable of travelling to Earth is vastly better equipped and if they're hostile, I'd think they could just wipe out Earth from a distance - but why would they travel from afar to do that? (Especially when we're so good at doing it ourselves )

Seems people is either "Aliens will try to vanquish us" or "Aliens could teach us a lot". Oh, and not to forget the third group saying "Aliens... We don't need no stinkin' aliens - we can F up our lives ourself"

I know you're kidding here, but I don't think we'll ever get a visit from any intelligent outer space alien lifeforms.

[Admin]

Carpet bombing  was ended because smart bombs is a cheaper way to reach a goal, with less protests from people against war.

And that's my pro-robot argument.

And the killer 'bot... How's it going to discriminate between a civilian and a soldier?
People wearing a green jacket and a hole punch may have been forced to do so and guerillas may look like civilians - how should it define who's who, when it's  sometimes impossible for humans?

The robot would need to be programmed with all the facilities/knowledge a human uses to make assessments. Obviously it is no simple task, and definitely not possible in the near future. But at some point I'd argue it will be. Maybe as early as 50 years time. Unlike humans, robots will be able to share information, employ sensors far superior to two biological eyes, etc.

If a robot intentionally kills a non-combatant, the current framework of laws that exist today will already consider that a war crime. An investigation will be made and the guilty party(s) dealt with, whether it be the programmer, the manufacturer, the politician, or the commander.

That said, I think the next tech to be fielded will actually be soldiers in exo-suits. But sooner or later, like aircraft today, the pilot will no longer be needed to do the job.

And no need to bash the US
(in every country there are good people, and there are bad people)

[Soeren]

Hi,

If a robot intentionally kills a non-combatant, [...]

That one is a no-brainer.
I'm more concerned with the unintentional behavior.
You may think it is possible to make the distinction under all ambient conditions, but I doubt that and then there's at least two ways for the enemy to render a robot useless: Either make some civilians look like combattants to a robot, while not to a human, get them killed and make an international case of it, which should make the controllers shut the 'bots down. Second, make the combattants look like civilians and the robots won't harm them. Either way, the robots will be as useful as wet towels in that scenario -  I think it was you who posted a link to how a drone was captured by fairly simple means. Humans are extrermely adaptable.

That said, I think the next tech to be fielded will actually be soldiers in exo-suits. But sooner or later, like aircraft today, the pilot will no longer be needed to do the job.

At least the research into exos will spill over to civilians, in turn helping people with a range of handicaps.

And no need to bash the US
(in every country there are good people, and there are bad people)

But it's OK to bash eg. Russia?

I could bash Denmark as well, as we were tainted by our former PM, who courted to be Bush's little lapdog, but luckily we had a change of government too, so finally we can get our troops pulled from Afganistan (where they shouldn't have been in the first place - althoug it isn't mentioned a lot, Denmark is considered at war because of that).

I know there's a huge number of good people in the US! I know a number of them and my GF's best friend married one. I have known quite a few good people from Iran, Iraq and Afganistan as well.

The problem with any country isn't really the citizens, but what their sometimes war mongering government can sometimes brainwash them with ideologies or threaten them to do their biddings.
Besides psychophats, sociopaths etc. nobody want to see other people killed (unless feeling threatened to a huge degree) and that's not due to ethics, but rather a function of the primitive brain, of as good as any species.

[Admin]

I'm more concerned with the unintentional behavior.

I'm confident robots will be tested sufficiently before being released. No one arms extremely dangerous weapons if they aren't darn sure it won't go berserk on them.

You may think it is possible to make the distinction under all ambient conditions, but I doubt that and then there's at least two ways for the enemy to render a robot useless: Either make some civilians look like combattants to a robot, while not to a human, get them killed and make an international case of it, which should make the controllers shut the 'bots down. Second, make the combattants look like civilians and the robots won't harm them.

Humans too can be fooled just as you described. I can think of a dozen examples. But again, my argument is that robots will not be used until they are shown to be at least as equally as capable as humans for discerning such things. It would be unethical, and a war crime, to use a weapon that cannot discriminate between combatants and civilians. Perhaps you have less confidence in robots in the far future than I do, but I claim that one day robots will be better able to discriminate between the two than human soldiers could, and in that case it would be unethical to not use robots. But hopefully we would have outgrown war by then!

But it's OK to bash eg. Russia?

If it makes you happy, you can bash Russia all you want

finally we can get our troops pulled from Afganistan (where they shouldn't have been in the first place

To be fair, the Afgan government had killed over 2000 Americans in one day.

But lets focus on only robots here . . .

[MrWizard]

Many people are scared of robots doing things were humans can not shut them down easily. So in my opinion there is still a long way where robots have to ' prove ' themselves as trustworthy autonomous beings.

But a robot can only be as good as the creator programs and invents them. So the human error factor will be included, certainly if the options list is getting bigger.

Makes any sense ?

[Gertlex]

So in my opinion there is still a long way where robots have to ' prove ' themselves as trustworthy autonomous beings.

But we don't require that of humans either