Ethics of Algorithms
Ethics of Algorithms
H/T Iblis Bane
Would you drive a self-driving car that would kill you in order to save 10 other lives?
#dataethics #ethics
Originally shared by Iblis Bane
The ethics of algorithms. Not something I'd ever much thought of in a practical sense, but on reflection, obviously something that needs to be examined now, not later.
Of course, there is a fundamental difference between telling the car what to do in certain situations, (imposing our own morality on it), and making it possible for the car to make that kind of decision itself.
We'll start off with the former of course, but the sheer number of variables that will arise make it impossible to pre-program every possible one. Inevitably, we must make it possible for the vehicle to apply one of a series of options to any individual event.
And this brings me to an objection to vehicle automation that I first learned from R.A Heinlein...the computer will follow its programming, regardless of the consequences. A human operator will at least try and salvage the situation, and human ingenuity has proven surprisingly successful at this in the past. The computer has no stake in the outcome.
(Ping Gideon Rosenblatt and David Amerland 'cause they may be interested. :D )
http://www.technologyreview.com/view/542626/why-self-driving-cars-must-be-programmed-to-kill/
H/T Iblis Bane
Would you drive a self-driving car that would kill you in order to save 10 other lives?
#dataethics #ethics
Originally shared by Iblis Bane
The ethics of algorithms. Not something I'd ever much thought of in a practical sense, but on reflection, obviously something that needs to be examined now, not later.
Of course, there is a fundamental difference between telling the car what to do in certain situations, (imposing our own morality on it), and making it possible for the car to make that kind of decision itself.
We'll start off with the former of course, but the sheer number of variables that will arise make it impossible to pre-program every possible one. Inevitably, we must make it possible for the vehicle to apply one of a series of options to any individual event.
And this brings me to an objection to vehicle automation that I first learned from R.A Heinlein...the computer will follow its programming, regardless of the consequences. A human operator will at least try and salvage the situation, and human ingenuity has proven surprisingly successful at this in the past. The computer has no stake in the outcome.
(Ping Gideon Rosenblatt and David Amerland 'cause they may be interested. :D )
http://www.technologyreview.com/view/542626/why-self-driving-cars-must-be-programmed-to-kill/
Glad you liked it Zara Altair. :D
ReplyDelete...and then we cross over into algo-bias as, how will "The Car" judge whom to save if I have an infant in the car ?!!
ReplyDeleteHart GA Infant in car! Yes. One of many questions.
ReplyDeleteThese ethical factors really need to be addresed right now. Thanks for sharing Iblis Bane Zara Altair
ReplyDeletePeter Hatherley They do! Thanks for reading.
ReplyDelete, since the SDC is issued a CA "drivers license" will it be allowed to cast a ballot?! '; )
ReplyDeleteThis dilemma was the underlying theme of the movie iRobot wasn't it?
ReplyDeleteWill Smith played a character where someone was saved by a robot while sacrificing someone else and Will Smith had a different opinion.
All very interesting debate and as the article pointed out very human reaction of "let harm come to them rather than me!"
Thanks for sharing Iblis Bane and Zara Altair
What an interesting article.
ReplyDelete