Can Morals Be Programmed?

I like the direction this thread is going. I don't think it is easy to delineate a rigorous palate of morals to code into a robot. A trivial claim might be "The greater moral action is the one that brings the least harm to the most people." The trolley problem for example suggests a distinction between switching the trolley to strike one instead of 4 is more easily contemplated than pushing one to his death to save the 4. It also indicates that pushing one to save 4 is more easily contemplated than carving up one living organ donor to save 4. I think the hard part is getting the robot to comprehend what formulates the basis of such codes when we can't come to terms about them ourselves. If the greater number were an absolute then why do we find the later notion so repugnant? I always find myself asking "What would a borg do?" And as alway I come to the inescapable conclusion that morality is contextual. We're we borg, hurting another would be as avoided as hurting yourself. Indeed the distinction is rather mute. Is it not the contextual scenario that man finds himself not as borg (for instance) that finds us in the types of moral dilemmas we contemplate?

I would suggest that we start by giving the machine a set of simpler tools like "Love", "Hate", "Guilt", "Remorse", "Empathy", etc. The system will needs these tools when it gets to the point of defining morals. Since humans can't define these terms accept on an individual basis (If That), my guess is that each machine will end up with a different but correct result. Could be an interesting experiment but not something that I would want to bet my life on!

Just My Thoughts

i give you a GA for "Could be an interesting experiment but not something that I would want to bet my life on!"

but i also think you are right about the emotional response criteria being needed.

While I said that Morals can and will be programmed... I'm going to venture out and say that you will not be able to program love any more than you can give the machine life. If it appears to be love, it is simply passing the Turing test by encoding the responses humans expect, and will not be genuine... because there is no spirit from which to have these responses spring forth from. This will be a much harder challenge than mere morals.

Chris

You give me the opportunity to close a loop here. One of the unexpected consequences of programming the moral code of the three laws of robotics is that independent of external programing a robot truly experiences love. Read Asimov's short story "Bicentennial Man" from the "I Robot" collection of short stories. They did make this into a movie of the same name, starring Robin Williams. While details were artistically added and omitted for the more visual media of a movie, the fundamental story was reasonably preserved in the movie.

Well, everyone has wondered way off-track. Not that doing so id bad, but the original question centered about "Do you think automated vehicles can apply the rules of war...?"

My take is that this has little or nothing to do about emotion, but can a machine interpret the rules and boundaries of war. These rules and limits are tangible things and therefore pretty straight forward to instill into a machine with today's technology.

That is quite a different thing than applying moral adjudication to life.

Ah, to kill or not to kill, that is the question!

GA.

The democracy of the Greeks is not equivalent to our present democratic systems. In many ways, the political system itself was better than ours, in so far as it was a better system for representative democracy. No professional politicians! Need I say more. However, the representation was exclusive to male citizens, and was not in line with present ethical and human rights standards.

The ethics of the ancient Greek democracies set them apart from the present questions. Women, and slaves, had no right to citizenship and played no part in the "democracy". The ancient greek ethic would have been OK with the idea of using a woman or a slave as a "human shield" - the practice you decried, Anonymous Hero, and used to distinguish our own warring and killing from that of the Taliban and such.

Our founding mothers fought for women to obtain the vote and to be recognized as persons under the law. This is a very recent step in our cultural evolution not a hundred years old. You have leaned on this legacy of our mothers to justify the war, and others have done so as well in the public eye. The exploitation of women and children as "shields" and "weapons" is repugnant to the majority of persons in this day and age, because we respect the same 'universal' code of human rights. The greeks and their version of 'democracy' did not.

Thanks, but the question was what is a moral killing machine, not what morals a killing machine has.

Now, what is moral? That is truly subjective and depends on what the group or society decides it is. That was my point about the SS Troupes. They are not my set of scruples, but that is me (and probably/hopefully most of us).

What is moral? What would you like it to be? Your choice, at least for you.

As for the IF THEN statements, it all depends on what level of abstractions you want to operate at. Software is getting better and better at operating at higher and higher levels of abstraction, too. Not long ago a computer could not challenge a master Go player (Wei Chi), but that has changed. Go is far more complex from a strategy point of view than chess. Chess can be won by brute force, but harder (not impossible) to apply brute force to Go.

However, many computer programs operate based on a best statistical analysis basis, so it isn't just yes/no decisions at a primal level. You can make weighted decisions and that is one of the algorithms autonomous missiles use in the game of predator/prey where spoofing and anti-spoofing are constantly in motion.

I do realize that this tangent on this tread was the question "What is a moral killing machine?" My second rebuttal point was intended directly toward that question. I have no idea which of humanity's moral codes the questioner is applying to this question. If he's applying the moral code that the killing of any life is wrong except for the purpose of producing food. Then the fight against pestilence and disease can be considered an amoral act and a butchery slaughter house would be considered a moral killing machine. If morality should instead only concern acts against people then Michael Vick should not have been charged with a crime let alone imprisoned. But because this amorphous code of morality cannot be perfectly codified by people to stand the test of time, doesn't mean that a machine could not get somebody's moral reference to work from.

I have to finish this now, the moral code of my Windows XP operating platform need to be updated.

Apologize for the tangent. My whole point in asking the question was that I do not believe that a machine can be programmed to know what morals are no matter how you may define them. A machine executes code. Humans may impart their incomplete definition to the code but the machine will never "UNDERSTAND". (At least in my lifetime)

Just my Thoughts!

All,

Please consider this - the article says, "the automation respects the rules of war." It does not say anything about programming morals into a computer or robot. Rules have a "1" or a "0". Therefore, rules can be programmed.

A CR4 Guest