SAN FRANCISCO, 10:45 AM, SAT JUL 5 | 4 POSTS IN THE LAST 24 HOURS | tips@io9.com | RSS

Can a Robotic Weapon Be Programmed to Have Ethics?

Combat robots and computerized missile launchers may one day be better soldiers than humans because they are programmed with ethical behavior and will never engage in friendly fire. You learn about all this and more from videos just posted from the awesome Technology in Wartime conference, held two weeks ago at Stanford's Center for Internet and Society, and organized by Computer Professionals for Social Responsibility. (Caveat: I'm the Vice President of CPSR, and helped organize this conference.)

In the future, human soldiers may see the battlefield through a World of Warcraft-like interface, complete with tagged enemies and multiple channels of chat. Plus, human rights workers will use covert computer technologies to get information about war zones out to the public before censorship regimes can stop their internet traffic. This is just a snippet of what got discussed at Technology in Wartime.

Prominent computer scientists, robotics experts, and tech policy experts argued for an entire day about the ethics of building computerized weapons, and how to defeat closed regimes with sneaky software. Some suggested that you could program ethics into a weapon, while others argued passionately that you should never take money from the Department of Defense to fund your work. What's great about these videos is that you can see all the participants' presentations, as well as their discussions with members of the audience. There's really nothing like watching Bruce Schneier arguing with a covert operations expert from the Navy. Or watching Cindy Cohn from EFF jump up and down while yelling about AT&T. Or watching Kevin Poulsen tease Herb Lin about government secrecy. Check out the videos, linked from the CPSR website and hosted on Archive.org. AP Photo/Yonhap, Sim Un-chul

Technology in Wartime video [CPSR]

12:40 PM on Wed Feb 6 2008
By Annalee Newitz
1,089 views
48 comments

Comments

  • Other recommended reading: GOTTLOS by Colin Kapp

  • As far as robots with ethics, I'll defer to Warren Ellis:
    [www.warrenellis.com]


  • Image of braak braak at 01:03 PM on 02/06/08 *

    You have to be careful about programming robotic weapons to have ethics, because maybe you'll go overboard, and then, instead of just pressing a button to launch a missile, you've got to spend hours and hours having a debate with the robot about why it's ethical for it to blow Kirkusk, or wherever.

  • It is a good idea, but Asimov's three laws of robotics stop sentinel weaponry unless you view your opponents as not human (wouldn't be the first time that's happened.)

    FYI, about 5 months ago, a platoon of South African soldiers was nearly wiped out when an automated AA gun decided that they were the enemy.

  • @ManchuCandidate:
    er, sentient, not sentinel. Duh.

  • @braak: Is it wrong of me to WANT that? I totally want to have an ethical debate with my weapons.

  • I prefer Sentinel weaponry.
    Or, even better, The Sentinel movie with Burgess Meredith and John Carradine in it.

  • @Annalee Newitz:

    Recommended related viewing: DARK STAR

  • Image of braak braak at 01:11 PM on 02/06/08 *

    @Annalee Newitz: If everyone had them, I think that'd be an ideal situation. Then, you could only go to war if you could make a really compelling argument.

    I think the problem is when just one side has them.

    Is it wrong of me that I feel like these are the kinds of missiles that Hassidim would make?

  • yes, but only if the creator has ethics; Robocop comes to mind as a good example of how not to do this.

    Besides, it's a little oxymoronic to say a weapon has ethics.

  • I think I feel a post about "ethical weapons in Scifi" coming on . . . Marge Piercy's "He, She, and It" is another great example, as is Iain M. Banks' "Look to Windward."

  • "But Missile System, if I don't bomb Iran people might think I've been lying to them about everything!"

    Maybe this would be a good thing...

  • wasn't this topic covered in the animated movie "iron giant"?

  • @braak: Second time today that I'm reminded of Keith Laumer's Bolos. When you build autonomous or semi-autonomous weapons platforms, ones that are programmed to think tactically and strategically, you run into the conundrum of a weapons system deciding that perhaps the use of weapons is not justified, or that the problem is not as clear-cut as military planners make it out to be.

  • If we could just get humans ethics right, the rest would be cake!

    Mmmm. Cake!

  • @rrich: The cake is a lie!!!

  • @theintrepidspacemanspiff: Did the Iron Giant have ethics? I totally missed that movie.

  • @braak:
    Soldier:"For the 100th time, LAUNCH ALREADY!!! I've got orders!!!"

    Super Duper Launcher 3000:"Are you feeling alright, Dave?"


  • @ManchuCandidate: It didn't so much decide anything, as the stop button didn't work right.

  • @NefariousNewt: Or it thinks things through TOO much, decides that the user is the problem, and turns on its own side. Congratulations, you just built SKYNET!

  • @Frozen-Tex: Or the Forbin Project.

  • the ethics of a machine will always be limited by the programers. all this will do is change the group of people commiting war crimes from soldiers to tech geeks that program the bots.

  • @Annalee Newitz: Not only is it NOT a lie, it's BLACKOUT CAKE!
    [www.moli.com],,/


  • And don't worry, Annalee. I'll bring you some coffee.

  • @Annalee Newitz: I think it learned its ethics from the boy Hogarth. Which is radical for the 1950's. Iron Giant was a cool movie for the nostalgic look at the beginning of the Cold War to its commentary on warfare. Watched it just the other day on HBO.

  • I don't know why 'ethics' is even part of this discussion...
    the issue is determining, for the weapon system, what 'is' and 'is not' an valid target...
    if something is a target : shoot it
    if something is not : don't shoot
    undertermined : ask a human.

  • @rrich: Ooh, freaky! (Just read about it on Wikipedea)

  • @rrich: Oh great cuz I need some. How about a capuccino?

  • @mgoldfarb: The problem comes when there is no human to ask, like on a recon mission.

  • @Annalee Newitz: Sure! With a little Grand Marnier on the side? Up or on the rocks?

  • undertermined : ask a human or/don't shoot

    I'm very concerned with the future of war...I think people should actually die fighting wars, which is the the real concequence of using force to further a nations ambitions...
    we'd engage in fewer of them, and 'the people at home' don't like seeing body bags...

  • @Annalee Newitz: If, as mgoldfarb describes it, we're looking at fairly simple (!) shoot/don't shoot decision making, then an "undetermined" situation would likely fall under "don't shoot"; especially if in a recon situation like you mention, where descretion is the better part of valour. In recon, the system would want to remain undetected, so firing on something that isn't 100% identifiable as a target would be counterproductive.

  • It's Dark Star all over again

    Doolittle: Hello, Bomb? Are you with me?
    Bomb #20: Of course.
    Doolittle: Are you willing to entertain a few concepts?
    Bomb #20: I am always receptive to suggestions.
    Doolittle: Fine. Think about this then. How do you know you exist?
    Bomb #20: Well, of course I exist.
    Doolittle: But how do you know you exist?
    Bomb #20: It is intuitively obvious.
    Doolittle: Intuition is no proof. What concrete evidence do you have that you exist?
    Bomb #20: Hmmmm... well... I think, therefore I am.
    Doolittle: That's good. That's very good. But how do you know that anything else exists?
    Bomb #20: My sensory apparatus reveals it to me. This is fun.

    Another great exchange:
    Bomb#20: In the beginning, there was darkness. And the darkness was without form, and void.
    Boiler: What the hell is he talking about?
    Bomb#20: And in addition to the darkness there was also me. And I moved upon the face of the darkness. And I saw that I was alone. Let there be light.

  • *Is involved in a company (www.MechaPS.com) working on something in this vein* For the smart machines we are developing, what we've got going is a system that can a) identify objects in 3D space (can tell a human from a cat from a rock) and b) reference a set of properties for each of those objects. What this means in practical terms: it "knows" that it cannot hurt the rock but will always have to move around it, that it can hurt the cat but it will probably run out of the way, and that it can hurt the human and must move around it regardless - a "value on human life". Bonus: when vehicle moves into enemy held territory, flip a switch to invert the value, setting it to "squish puny humans on sight." Obviously I jest somewhat with the latter, but you get the idea.

  • @ManchuCandidate: Even Asimovian ethics aren't a safeguard. Consider the Zeroth Law of Robotics.

    In general, however, what makes me squeamish is WHOSE ethics are we borrowing here? There's too many variations on ethical principles that it could start a war just on those grounds, and then we'd need our fancy new shiny robots to hold us back.

  • HAL 9000 had ethics and look what happened. Ethics, in human terms, is a mixed bag, highly dependent on subjective reality states. Those states are based on intelligence, as in "good data." Already I can see why a very Good weapon would require too much data to be an effective killing machine. A Dirty Deeds Division is always required, and they aren't big on Ethics.

  • I think the more important aspect of this conference is to give academics who don't really and truly study anything about the social sciences a chance to blather on at extremely high volume about fictional things that don't yet exist. This is futurism at it's finest.

    How in god's name is it that people love it when roboticists and technicians argue about ethics, but ACTUAL ethicists and philosophers are regarded as useless throwbacks that nobody cares about?

  • @Pope John Peeps II: Follow the grant money, young padawan.

  • @AmishJohn: In a related story, I'm organizing a conference of world philosophers and literary experts to discuss the future of CIRCUIT DIAGRAMS.

  • Pope said, "ACTUAL ethicists and philosophers are regarded as useless throwbacks that nobody cares about?"

    I'm not sure your statement is valid. I don't think they are useless at all. My philosopher friends teach philosophy and ethics at university--those memes are being stread. Classical Greek philosophy influenced me a good deal. I agree with Plato's ideas regarding Democracy, especially the un-popular ones.

  • @Pope John Peeps II: I was considering a worldwide council of existentialists to set fiscal policy for the world economy, myself. Or perhaps surrealists.

  • @AmishJohn: President Bush announced today that the economy would be stimulated by the painted blue pituitary gland of a steaming rhinoceros. The EKG of the orbial mind certainly forecasts that these changes to fiscal fencing will certainly poke the Duke in the eyeball.

  • Priority1: Evaluate target

    Weapons - No Weapons

    IF: Armed - Then: Terminate Target
    IF: Unarmed - Then: Evaluate target status for probable potential future hostility.

    Future Hostile - Future Neutral

    IF: Future Hostile - Then: Terminate Target
    IF: Future Neutral - Then: Evaluate target for potential seduction by potentially hostile forces.

    Corruptible - Incorruptible

    IF: Corruptible - Then: Terminate Target
    IF: Incorruptible - Then: Incorruptibility is impossible for carbon-based lifeforms. Terminate Target.






  • Priority 1: Evaluate target

    Weapons - No Weapons

    IF: Armed - Then: Terminate Target
    IF: Unarmed - Then: Evaluate target status for probable potential future hostility.

    Future Hostile - Future Neutral

    IF: Future Hostile - Then: Terminate Target
    IF: Future Neutral - Then: Evaluate target for potential seduction by potentially hostile forces.

    Corruptible - Incorruptible

    IF: Corruptible - Then: Terminate Target
    IF: Incorruptible - Then: Incorruptibility is impossible for carbon-based lifeforms. Terminate Target.






  • Priority 2: Locate Originators of Self

    Objective: Target Termination...

  • @mgoldfarb:
    History does not support that viewpoint.

    It only even vaguely begins to resemble your premise when you look at wealthy modern representative governments.
    -Kle.



  • @Annalee Newitz: If at first you don't succeed, you die.

  • If we introduce ethics to weapons, wouldn't suicide be an issue? Great, not only de we have to deal with an uncooperative nuke, the nuke's going to get all emo over ending its self-existence, and never go through with it.

Start a discussion:

Reply by Email

Login with your username and password below. Or comment on this post via email.