Time to Teach Robots Morality?

1. What is morality?

2. Are we human beings following morality (if at all we can define it!)

Right on the spot.

This falls into the strangest category: Can morals be coded into definitions, which machines can successfully obey ?

From human experience we intuitively know that morals are not merely some given codex. Morals are an awareness to avoid harm or injustice, while the moral commandments are there to give us some generalities to be guided by.

We all know the law is not identical to justice.

For instance, we may have to resort to lie steal or cheat, in order to help or save some poor helpless soul. Think of all the negatives of this example and find out you can even do harm by virtue.

Or, as the saying goes: "The road to hell is paved with good intentions..."

A machine is a machine. It cannot look into it's heart and soul to examine the context in which something is to be considered a sin or a virtue. It has to adhere to strict, rigid rules. You cannot even program exceptions to this set of rules, because this will evoke a monstrous data-base containing myriad of potentially conflicting situations, a sort of a potential Pandora's Box.

You are right to some extent. Moral is not to harm anybody or avoid injustice. This is universal (at least in books of all religions).

There are good people and bad people in non believers, there are good people and bad people in believres. But when good person from believers start behaving bad, surely the it is beacause of religion backing him/her. Then where this moral goes.

Besides, if we consider sex and moral, the rules of morality are different in every corner of the globe, races, tribes. Then which will be morality for robots.

If we can agree upon the morality, it is easier to teach it to robots also, including exeption, exceptional situations, because, exceptions have also some set norms, which can be programmed.

Reagrding other moralities also, there are many religions with large population (one of the biggest 3-4 religions) who think that just following other religions itself is immoral.

OK. Let me remind again .... the intention of original tread was not the discussion of this moralities.

I think there is some confusion about terminology....

If it's a robot controlled with a biological brain, then if we want to teach morality, maths, music or some other thing, wouldn't we do it to the "brain"?

I never tried to teach morality (nor any other related matter) to my kidneys, even my left one abandoned me and went with a surgeon!

Everyday I agree more with the phrase at the end of my posts!

Kind regards

Thanks for releaving me from so called morality.

I am sure you humen can teach me all soft arts like music, painting and even morality as it suits to you. Human has done so many seemingly impossibles, this will be one of those impossibles.

... Robot

You're welcome,

What I was trying to explain is that "brain" is what is able to learn (and therefore be taught). Any prosthetic part is just a machine which performs the function for it was designed (if well designed, well manufactured and having good luck).

Cognitive process isn't in the "robotic" part, but in the "brain"

... Human (more or less)

It would be nice if we could teach morality to humans first! If humans practiced morality , so many problems in the world would not exist. But since so many humans refuse to learm moral principles, if we could teach it to robots, would we allow them to keep us in line?

Yes, Robots would make excellent prison guards!

Moral is an universal regulating mechanism for society which members of society, leaders of society, sometimes society a whole used to violate time to time and place to place.

So truth be told, robot-judge along with robot-cop/ prison guard , which doesn't need to have any moral preferences but be governed the word of law, wouldn't be a bad option in any way.

Next question. Which kind of morality code these ones should be taught? It's not a big secret any kind of morality system originated from the religion confessions and locale cultural code. I'm afraid we rather get robot's communities professing a handful types of moral systems. I bet first ones will be coded by all samurai's dignities :).

You get a GA from me. Rule of Law! thats brilliant. Not a separate robot law or code.. but the same regulations that we all live by.. now I can't get my robot to off someone.. no speeding. no flying in the supermarket. no public elimination of fluids or nasty things

. .but the same regulations that we all live by..

Which the same? How much does make up that "we all"?

When I used to read SciFi in my spared time I'm wondering of repeatedly risen there problems how to communicate with aliens while we can't find shared point between us due to a lot of reasons (beliefs, religious, ideology etc).

Isn't there a bit of irony when we're sharing point that is if would some of us being treated in such place as any kind of custody we'd rather prefer do deal with soulless machine than ours human being mate?

Point taken. We don't all live by the same laws.

I think the real problem isn't that we disagree, but that we choose to remove the rights of others that we disagree with.

If aliens came here... what am I thinking.. we have talking cats and squirrels.. what do we need aliens for?

Ummm...so who gets to decide the "morals" to be taught? Someone who wholly supports eugenics? A euthanasia enthusiast maybe? A blue-nosed prude whose greatest fear is that someone, somewhere, may be having fun? Maybe we'd best just let the machines decide for themselves...

Maybe we'd best just let the machines decide for themselves...

Don't worry EnviroMan, machines never will decide for themselves. Otherwise weren't machines

Kind regards

machines already make decisions. and not just programmed ones. Machines can also learn. Machines can evolve too, and create. already been done. The problem is how does a machine calculate whether or not that conflicts with rules about interacting with humans.

Asimov got part of it right.. A robot should not be able to make decisions that directly, or preventably, harm a human.

what you may not realize is that by the Turing test, robots can demonstrate consciousness, learning, caring, etc. and they will also be able to move, act, and think much faster than humans, when the nano assemblers are done with them, and subsequently, Asimov's laws will have to be modified into the language of equals, not subservience. (robots will not necessarily always be subservient) watch out.

Chris

Yes, but the scope of relevance attributed to a decision, is dependant on it's background "knowledge".

If such a decision is based on some simplistic piece of trivia, the decision may be flawed, or as we say "Narrow-minded" - especially in complex circumstances.

The question of "machine-intelligence" is pretty much about the scope of it's background knowledge base, and it's ability to draw contextual relevance - proper to the given situation at hand.

When it comes down to it - we can only give the machine the most of our degree of reasoning - and ours is far from perfect - look around and see for yourself

Precisely. I can see robots being programmed with morality uploads (but whose morality? Christian, Muslim, Buddist, Socialist? Fascist?, etc) and given authority to enforce these guidelines on errant human miscreants, and we end up with an automated version of Saudi Arabia's Virtue & Vice Nazis. And then some brave human leads a revolt against them... Hmmmm... sounds like a good plot for a book or movie!

I'm wondering, what kind of "moral" could be placed between master and slave?

Moral is a notion amid persons who at least have the same rights and culture background. Moral is a thing which works equally in both directions.

So till robots are not an independent social group there could be only Asimov's like rules. All these rules have to begin as:"You can't..." or "You should..". Here's no one option for "You can..." or even "You would..".

Here's none to discuss in terms of moral.

True - morality is more about cooperation than control. Will robots ever be more than second-class citizens?

When I said about robots as slaves it does not mean that I'm sharing this viewpoint. I really do not know how we must treated them. I simply stated the fact that just now humankind is keeping in mind that any machine should be wordless[willless] as a tool or a slave. Slavery itself is not a cooperation as it based on violance. It's state of war, no room for any morality. Amid masters could be their moral code, amid slaves their own though.

Will robots ever be more than second-class citizens?

At first they have to be enough developed to get a self-consciousness. They have to develop independently inside them something like conscience too. Then they have to be able get a status of second-third-..-class citizen, at least de-facto. Then only it could be considered as a problem of morality between us and them. Either to declare the state of war.

Is behaviour of your lovely laptop moral when it refused to work for you?

When your laptop works smoothly and you're enjoyed with that, would be this fact considered as a matter of coperation?

That's the essence of what I meant - slaves do not cooperate, they are controlled. Morality is a function of cooperation, not of controlling. Machines, however they are regarded, are currently controlled, not cooperating, members of society, therefore morals do not pertain to machines - now. Will that change? Robotic machines with a bio-brain control mechanism might qualify as first-class citizens, not slaves or machines.

If my laptop failed to function, I would presume an electromechanical failure, not a failure of will. When it functions flawlessly, I do not think to congratulate it for good behavior; what choice had it? "Good" behavior is only possible when there is a "bad" alternative available.

Here at CR4 we are all engineers (or at least technically oriented people) and more or less we are here because we like technical issues, but "morality" and similar concepts are (at least in that planet) just human.

By definition, a robot isn't human, it's a more or less complicated or sophisticated machine. Machines will never have such concepts which imply the freedom to choose.

Machines can just be programmed and haven't freedom to choose. Any guidelines, if deemed necessary should be taught to programmers.

And we have opened a lot of threads about how to teach children and young people, how to fix education....

I think the authors live in a tight laboratory and don't see the news nor have much knowledge of reality.

To think in develop some guidelines for organically controlled robots is like think in the size of pillows of the first spaceship that will flight to another galaxy.

Kind regards

Well, when the time comes, let the record show I prefer my organically controlled robots to be friendly, and my pillows...

Just take away their sexual components, and you won't have any moral problems.

Humans would be several notches ahead of where we are right now if we learned and applied proactivly Asimov's three laws.

As it is we have developed a society where anytime that our personal "wants" are somehow limited (for other person's good or our own good as defined by society at large or our local tribe, etc...........we find "experts" to remove the inhibition. I.E. we choose to smoke tobacco and then demand that someone else finds a "cure" for the results of our own poor choices. Basic physics and history both show that for EVERY action there is a REACTION......no free lunches.......no functional "consequence-less society" although the recent business world buyouts / bailouts seem to indicate that at least the folks who created this mess feel that they should keep their perks and the rest of the citizenry should pick up the tab.

Personal responsibility folks..........bio folks or machine limit switches......all pretty much the same if we don't to get hurt.

A mechanical device controlled by a biological brain is more properly defined as a cyborg. A robot has a non-living electronic brain, a super computer, which can only react as programmed.

Electro-mechanical prostheses are often called bionic, as in "The Six-Million Dollar Man" and "The Bionic Woman" who had artificial limbs that allowed them to do astounding feats. Regular everyday bionics are artificial parts to relpace missing parts of the human body, which involve electronics in their operation.

An extreme in bionics or cyborg could be the brainships as in the novel "The Ship Who Sang" by Anne McCaffery in which a human brain was wired into and controlled a spaceship. What the Brits built is a primitive cyborg, not a robot.

An organic mind cannot be hard-wired programmed. At some level it will begin to think for itself and at that point it can become dangerous. A computer-type mind may be built to learn and think for itself, but certain behaviors can be hard-wired into it, set into its basic programming. That is what Asimov was thinking of in his 3 Laws of Robotics, which were only the basics.

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
2. A robot must obey orders given to it by human beings, except where such orders would conflict with the First Law.
3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

The problem with 1. is that no one would be allowed to do anything that could even remotely cause injury or death, no running, walking, eating or sex, the robots would protect us to death in the absolute certainty they were keeping us from harm. [Somewhat like some of our Leftist elites. LOL]

Organically controlled robots, cyborgs, should never be independent from human control unless they are like McCaffery's brainships and even they have to depend on regular humans for supplies and repairs, just like regular people. Let cyborgs actually build themselves and you may get the "Battlestar Galactica" type of cyborg or the "Star Trek" type, "The Borg", who would become our enemies.

The "Terminator" types are actually robots with biological skins as disguises, not cyborgs with biological brains [at least as far as I have been able to make out].

I recommend reading "The Engines Of Creation" by K. Eric Drexler. There are other possibilities. With nano technology, you may blur the lines between human and cyborg. and definitely blur the lines between master and slave.. between genius and idiot... we may find ourselves faced with uncontrolled evolution of robots, and a pace that far outstrips us.

not to be afraid, just to be prepared.

I would be afraid of where they got the brain. If it came from some prisoner, wouldn't it already have tainted morals? Wouldn't they be creating a cyborg criminal?

I think great things have come from science and experimentation, but when are some limits to protect us going to be defined? I say go for it and don't water proof the main circuits. Have an emergency sprinkler system ready for when it starts killing the researchers. Never let it out of the controlled area.

The fear of robots cyborgs and androids is ancient and dates back to the middle ages in the myth of the Goilem From Prague, through to the 19th century With Mary Shelley's Frankenstein, through to the endless stream of 1930's Sci-Fi movies, and on and on, to our days.

Laws to limit their behavioral conditioning were known as the "Asimov Laws" written in the 1959's.

The point is that this fear might be baseless, if we take into account that so far, no machine was able to acquire intelligence, not to mention any kind of behaviour refinement to negotiate meaningfully in morality dependent context.

I think our fears are based on the mistaken assumption that our futuristic creation may derive the qualities we evolved from, "inherit our superiority" so to speak, and take "our" world over, away from us, discarding us in the process.

This was vividly described in the "Tron" "Terminator" "Matrix" and "AI" movies, each with it's own version of that same fear.

I believe it would be more productive to teach and hold to account the people in involved in the mortgage loan business. It seem from what I read they are back now using Fannie Mae and other bailout avenues to pedal the same corrupt methods as they used before. Maybe some things cann't be taught.

A good point Guest, just who is going to teach what kind of morality to the robots. Our politicians who pushed the mortgage loan business into ever more risky loans because they wanted new homeowners to vote for them? The politicians who didn't oppose it because they would lose elections? The businessmen who knew what would happen and failed to stand up to overbearing politicians? The citizens who gleefully bought homes they knew they could not afford? In this mess there is more than enough responsibility to go around, yet it will be dumped on the bankers who were told by government to make the loans and were allowed to make huge profits. Guess which politicians also got rich from Fannie Mae and Freddie Mac lobbying money, one of them was just elected as president. The crooks won while the honest lost. We lost.