1. Published on Amazon? If you have a book, e-book, or audiobook available on Amazon.com, we'll promote it on WritingForums.org for free. Simply add your book to our Member Publications section. Add your book here or read the full announcement.
    Dismiss Notice
  2. Dismiss Notice
Thread Status:
Not open for further replies.
  1. Jack Asher

    Jack Asher Wildly experimental Contributor

    Joined:
    Sep 4, 2013
    Messages:
    3,571
    Likes Received:
    2,061
    Location:
    Denver

    AI ethical dilema

    Discussion in 'Debate Room' started by Jack Asher, May 21, 2014.

    From Erik Sofge: Should a robotic car be programed to kill it's driver in a crash if the alternative is to kill TWO people in another car?

    Presumably the cars are able to communicate and both are aware that a crash is unavoidable.
     
  2. thirdwind

    thirdwind Contributing Member Contest Administrator Reviewer Contributor

    Joined:
    Jul 17, 2008
    Messages:
    7,442
    Likes Received:
    2,938
    Location:
    Boston
    No. What if the two drivers are each over 90 and close to death and the one driver is a young woman in her prime? What if the one person is pregnant?
     
    obsidian_cicatrix likes this.
  3. obsidian_cicatrix

    obsidian_cicatrix I ink, therefore I am. Contributor

    Joined:
    Jul 15, 2013
    Messages:
    1,785
    Likes Received:
    1,557
    Location:
    Belfast, Northern Ireland
    Exactly the tack I was about to take. How does a robot assign value to life? And more to the point, should it be allowed to?
     
  4. 123456789

    123456789 Contributing Member Contributor

    Joined:
    Jan 28, 2012
    Messages:
    6,982
    Likes Received:
    3,630
    "It happens quickly—more quickly than you, being human, can fully process.

    A front tire blows, and your autonomous SUV swerves. But rather than veering left, into the opposing lane of traffic, the robotic vehicle steers right. Brakes engage, the system tries to correct itself, but there’s too much momentum. Like a cornball stunt in a bad action movie, you are over the cliff, in free fall."


    I'm not quite sure how this scenario constitutes as a definitive ultimatum. Miss ethics professor is assuming that the AI can't avoid a mishap if it goes right, and she is also assuming the person will absolutely survive if it veers left. What if the bigger car smashes into the little car, then the truck behind it smashes it into both of them? What if the person in the AI car gets unlucky and loses half his face, then commits suicide anyway, but only after having to see the look on his wife and children's faces?

    Better ethics question: Should people not smart enough to come up with technology themselves be allowed to raise pointless hypotheticals about it?
     
  5. Jack Asher

    Jack Asher Wildly experimental Contributor

    Joined:
    Sep 4, 2013
    Messages:
    3,571
    Likes Received:
    2,061
    Location:
    Denver
    Whether or not the person will survive any action is unimportant. The question is how is the car supposed to evaluate it's risk options when (not if) it comes time to program these cars.
     
    123456789 likes this.
  6. Lea`Brooks

    Lea`Brooks Contributing Member Contributor

    Joined:
    May 11, 2013
    Messages:
    2,892
    Likes Received:
    1,905
    Location:
    Virginia, United States
    It's like I, Robot! The thought of cars making their own decisions gives me the willies.
     
  7. stevesh

    stevesh Banned Contributor

    Joined:
    Mar 17, 2008
    Messages:
    968
    Likes Received:
    648
    Location:
    Mid-Michigan USA
    And this is the reason the lawyers will smother the autonomous vehicle industry in its crib. Not necessarily a bad thing, sez me.
     
  8. Aled James Taylor

    Aled James Taylor Contributing Member Contributor

    Joined:
    Sep 7, 2013
    Messages:
    797
    Likes Received:
    473
    Location:
    UK
    If the AI systems in each car can communicate with each other, this is not a single AI system making a decision concerning one car but the combination of two systems making one decision concerning how to minimize casualties in the inevitable accident.

    If the robotic cars are driving so badly that they come anywhere near having an unavoidable accident with each other, their AI systems are inadequate for basic driving decisions let alone ethical ones.
     
    Okon and 123456789 like this.
  9. Cogito

    Cogito Former Mod, Retired Supporter Contributor

    Joined:
    May 19, 2007
    Messages:
    35,953
    Likes Received:
    2,072
    Location:
    Massachusetts, USA
    Lawyers will NEVER smother a technology that will be earn them millions in litigation income.
     
  10. Burlbird

    Burlbird Contributing Member Contributor

    Joined:
    Dec 29, 2011
    Messages:
    978
    Likes Received:
    295
    Location:
    Somewhere Else
    Why do I get the impression that the main programming ethical dillema of any near-future traffic AI will have more to do with the insurance companies than with pregnant women and wise old men :)
     
Thread Status:
Not open for further replies.

Share This Page