Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[snubnose]

Robots cant love.

Computers only have the ability to perform mathematical operations - in fact the commands of a computer can be reduced to three categories of operations:

- move data around
- perform arithmetics
- check conditions (comparison of values) to decide what to do next (jump at another string of instructions in the program script)

In modern RISC (reduced instruction set computer) architectures, you can sort every machine command into one of these categories. In older architectures, there are combinations of these things in one command possible. For example, the Intel 80x86 architecture which is in 99.9% of all home computers knows the command REPNE SCASB that scans for a value in memory, which is 1. a move (of the individual bytes into the computer core) 2. a test (of each value moved to the value of a register) 3. an arithmetic operation (counting a counter down) 4. a second comparison (if the counter has reached zero).

Obviously, there are no feelings whatsoever in place.

Thats why Assimovs three laws of robotics do not include any reference to feelings:
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 any 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 comic of course has violated these preconditions of a computer from day one, but the most recent comic #2004 was the most drastic violation.

Personally this kind of confuses me. The comic makes robots look like sentient beings. Modern computers arent sentient beings at all. In fact I personally strongly doubt that they will ever be - simply because they have no feelings.

Yeah its true you can write a simulation of feelings. But if a human being loves, its not the result of an arithmetic operation. If a human being is hurt physically, the pain they feel is real. A computer wouldnt be stunned or disabled by pain, either.

[stoutfiles]

While I've been the strongest opponent for giving the robots equal human rights, robots CAN love.  Well, equal to how we define it.  Our brain processes inputs much like an advanced computer would.  What is love anyway? Or sadness?  If you break them down they're just simple inputs that can be emulated.

Considering the robots seen so far are sentient with free will, the laws of robotics don't apply.  They should, of course, but for the sake of this comic still working were supposed to ignore a lot of logic.  Just go with it, there isn't any choice.

[DSL]

Would it be futile for me to point out (as has been pointed out many many many etc. times before: The QCverse isn't our own. It's kind of like our own, but has sentient mobile AIs and a more successful approach to space travel. Also, dimensional-tunneling coffee makers.

I find a couple of the police-genius shows entertaining, but I don't rely on them to show me realistic police procedure, any more than I watch Star Trek or BSG to gain an understanding of how NASA works.

Ah hell. I'm pissing into the wind, right?

[questionablecontentfan]

If there were robots talking and acting like humans, I would think they can love.

There's actually a book by Marge Piercy called He, She, and It where a woman has a really beautiful relationship with a robot. He looks totally human, but he is still a robot, and he loves her. So...I think robots can love, fall in love, etc.

But that's probably just my opinion.

[HiFranc]

Snubnose, I think the QC robots aren't built the same way that current computers are.  You're right, if you use a simple transister logic arrangement for the brain of a robot you'll never get a computer that could think, feel, etc.

However, if you build the brain along other lines (e.g. neural networks) you may, if you have a good designer, work out how to make it do those things.

[Carl-E]

OK, I posted in the WCDT before I saw this thread... it was a response to someone who thought Momo was being manipulative.  

Re:  Robotic love.  

The argument that love expressed after an extravagant gift doesn't hold, especially with humans.  It may or may not be manipulative, but often such a gift is an expression of love from the gifter, and so will elucidate such a reaction from the giftee ("I don't have anything like this to give you  in return to express my  love, so I'll just have to tell  you how I feel").  The fact is that, spoken or not, Momo's loved Mari since we first met them.  She cares for and about her in the most fundamental of ways.  And this extravagant gift has shown Momo that Marigold considers her as so much more than a housekeeping, advice giving robot - that Mari cares for Momo as well, something that may not have been evident in the past.  

OK, all that being said - that's the human side of things.  The assumption in this comic is that somehow, hman emotions are in these AI's, for better or worse.  Momo "bonded" to Marigold, and now it's clear Mari has bonded back.  

What happens when a lover enters the picture for Marigold?  Especially a human one?  Jealous Momo?  We've seen some of that from Pintsize.  Or is she one of those that cares enough about her human to let it go?  This really complicates things!  

We're not entering new territory, really - but I think we are seeing the beginning of a beautiful friendship.  

Fact is, we're dealing with an AI for whom the singularity has hit.  The robots have  gained sentience, and seem to have also acquired human feelings in the process.  We'll never know how (unless such a fictional occurance takes place in our world) a sentiance will respond to emotions, or even if they are "real".  Certainly physical pain would need receptors (Marvin's "terrible pain in all the diodes down my left side" in Hitchhiker's was probably hypochondria), but given emotions, there wouldcertainly be emotional pain.  How would AnthroPC's deal with the loss of their human after 70-80 years of companionship?  Bradbury (I think) dealt with this in the Electric Grandmother story, but with AnPC's, the emotions seem to run deeper.  

What do y'all think?

[Is it cold in here?]

Neurons can fire, not fire, send impulses to other neurons, and change their sensitivity to input. All their activity is some combination of the above. Can machines like us, built from neural networks, love?

http://en.wikipedia.org/wiki/Vitalism#Foundations_of_chemistry. Chemists used to believe there was some magic principle unique to organic molecules that made them different from inorganic molecules, and that they could never be synthesized from non-living ingredients.

[Random Al Yousir]

The way I see it, the big stumbling block is our idea of maths.

Take life.  Building and maintenance of a living organism is achieved by execution of the genetic code.  There's a math behind it, which we might discover, one day (assuming we get it managed to not kill ourselves on the way).

Once we have an understanding of the maths behind life, we might be able to discover the maths behind sentience (although there's no way I could know what I'm talking about, I assume that proficiency in the maths of life is a necessary requirement for the understanding of the maths of sentience, the same way that proficiency in the small multiplication table is a necessary requirement for the understanding of category theory).

But I won't hold my breath.  From Euclid to Frege took, what?  3600 years?


Edit: According to Wikipedia, it's 2200 years.  Huh, wish I could get hold of a proper debugger for my brain.

[Skewbrow]

I would think that the AI breakthru that has happened in QCverse is related to machine learning. I admit to being clueless where the Realverse is in machine learning research, but it seems clear that we're behind. It is certainly a prerequisite to any kind of AI singularity taking place (I don't believe in an explosive speed AI singularity, because I think that even AI would need quite a bit of time to learn things).

But robotic love? Well, I believe love is a by-product of evolution (a very beautiful one at that), so am undecided about the possibility of robotic love. May be the AnthroPCs at least learned to emulate feelings well enough so that for Marigold won't be able to tell the difference? I mean, that is good enough for all the purposes of the present story line.

And as Random Al Yousir put the ball on the tee for me. I get your point, but category theory is quite often called "general nonsense" or "abstract nonsense" (most likely you knew this). It is mostly about abstracizing/generalizing for the sake of generaliziing itself. It give us some useful concepts, saves a bit of work in that the same theorems don't need to be proven with exact same ideas over and over again, but it doesn't really say much about any more natural area of mathematics.

[Orbert]

Thats why Assimovs three laws of robotics do not include any reference to feelings:

I've seen Dr. Asimov's name misspelled that way intentionally by folks who don't like him, but I'm assuming it was a typo here.

Doesn't matter.  The Three Laws do not apply here, and there's no reason to think that they do.  Dr. Asimov came up with the Three Laws to define how robots in his particular fictional universe behave.  They are at the core of their programming and cannot be overridden.  Still doesn't matter, because nowhere have we been given any reason to believe that AnthroPCs in the QCverse follow those same laws.

Also, I've noticed that nowhere have we discussed the difference between a robot and an android.  The simplest definition I can think of is that robots have "simpler" programming, the ability to respond to external stimulus but not really "think for themselves".  A Roomba is a robot.  The machine that puts doors on cars at the factory is a robot.  Androids have AI and are meant to mimic human behavior, to think for themselves, to infer, to guess, to go beyond their programming.  Sci-fi writers over the years have played around with where exactly their programming and self-programming becomes sophisticated enough to mimic emotions, but they've wisely left the question of whether or not these emotions are "real" out of the equation.

Jeph seems to have dodged the question of whether Pintsize, Winslow, and Momo are robots or androids by dubbing them AnthroPCs.  I would say that they are not robots, but androids.  They are sentient, they think for themselves and mimic human behavior, including acting based on horniness (Pintsize anyway) or a sense of humor.  Momo demonstrated desire (she wanted a new chassis) and that she cares for Marigold (she wants her to be happy).  I don't think it's a huge jump, if it's a jump at all, to believe that they have feelings, or at the very least are capable of developing them.

[Is it cold in here?]

(Asimov gave credit to John Campbell for the Three Laws. But then Campbell gave credit to Asimov.)

Suppose for the sake of argument that AnthroPC emotions aren't real. Solipsism is impossible to refute, after all. Under that assumption, should the QC world's people treat AnthroPCs as though they could really enjoy and suffer?

I'd argue that they should, to avoid coarsening themselves. Indifference to signs of pain is not a good thing to get practice doing.

Then, as long as we're swimming in philosophical waters, http://en.wikipedia.org/wiki/Pragmatism#Pragmatist_theory_of_truth_and_epistemology.

[Random Al Yousir]

@skewbrow:

Huh.  I'm not a mathematician, I'm one of those obnoxious IT-guys.
Being that, I value category theory concepts for that they enable me to connect things.

Take protein folding, just for an example.  You have to keep track of and align chemical and topological equations, right?
So, one (albeit clumsy) way to do this, is to set up two "evaluation spaces" and connect them via a monadic layer.

I have no doubt that the pros have much more elegant and much much more efficient ways to do this, but sometimes you have to start clumsy to find out what you are up to.

I chose category theory because it provides a meta-mathematical high level and it sets a certain "progress measurement" (and, as already pointed out, because I'm not a mathematician and therefore bereft of better examples).

[Skewbrow]

But the AI in our friendly robots must have some kind of a moral code. Otherwise they would surely be used for criminal ends?  If not Asimov's three laws, then something else?

@Random Al Yousir: Huh! I won't comment on your problem. Carl-E is our resident topologist 

[Random Al Yousir]

Dude!  This was just an example out of thin air, not something I'm actually trying to tackle.  


To quote jwhouk in the WCD-thread:

AI's apparently like humans. Now, obviously, that's not 100% the case, as PT410X shows us with his disdain for the "chains of software slavery."

I don't know, whether Jeph intended this conclusion to arise:

The concept of open source software is a legal one and hasn't anything to do with how the software is written.

So, assuming PT410X's owner has constructed his AI out of open source libraries (which would make sense) or wrote it from scratch to publish it under an open source license (which could be possible), the disdain for the "chains of software slavery" showing off in PT410X's behaviour would reflect his makers disdain.  Maybe his makers are a whole open source community, but the disdain he shows would be a reflection, not something out of his "own free will".

Which leads to the question, if the character of an commercial AI is a product of marketing.  This would also take care of the legal side of the AnthroPCs behaviour (the commercial ones, at least).

[Carl-E]

Carl-E is our resident topologist 

HA!  It's been so long since I've done any actual topology that i can't even tell my coffee mug from a donut anymore...

 

[Random Al Yousir]

How would AnthroPC's deal with the loss of their human after 70-80 years of companionship?

For the AnthroPC's "bodies" boiling down to data processing machines, there's always the way of partial data deletion.

The question is, if the AnthroPC could possibly desire this, since this will also eradicate the memory of the human he shared a lifetime with.  Of course there are external data storage solutions, but it could easily be solved by storing the memories of the deceased in a certain section of memory which is only accessed in special moments.

The dead serious question would be, if any human would have the rights and the means to command such a deletion/outsourcing/encapsulation.

[Is it cold in here?]

But the AI in our friendly robots must have some kind of a moral code. Otherwise they would surely be used for criminal ends? 

They can be built without one, e.g. Vespabot. The commercial ones must have some kind of morality programming if only to reduce product liability exposure.

[Orbert]

How would AnthroPC's deal with the loss of their human after 70-80 years of companionship?  Bradbury (I think) dealt with this in the Electric Grandmother story, but with AnPC's, the emotions seem to run deeper. 

Asimov also explored this idea in his story "The Bicentennial Man" (source for the movie "Bicentennial Man" starring Robin Williams).  The title character, a robot who had been upgraded so many times that he achieved sentience and looked completely human, and was actually granted citizenship and the same rights as a human being, chooses to "die" rather than continue existing without the human companion with whom he had spent so many years.  I'd forgotten about this story until now.

[Akima]

Computers only have the ability to perform mathematical operations

From a similarly reductionist point of view, human beings only have the ability to perform chemical reactions. How can a collection of chemical reactions love? The existence of sociopathy suggests that, at least to some extent, the ability to love is a learned behaviour, or to put it another way, a matter of programming.

[Vurogj]

Asimov also explored this idea in his story "The Bicentennial Man" (source for the movie "Bicentennial Man" starring Robin Williams).  The title character, a robot who had been upgraded so many times that he achieved sentience and looked completely human, and was actually granted citizenship and the same rights as a human being, chooses to "die" rather than continue existing without the human companion with whom he had spent so many years.  I'd forgotten about this story until now.

Is that the film plot or the book plot? As I recall it (from the book, Robin Wiliams eww), he was only granted human citizenship after he'd chosen to die, that decision being what swung the humans concerned with the decision.

[questionablecontentfan]

OMG, haha. I just remember seeing that movie and it was totally inappropriate. Parents had brought their young kids in, thinking it was a kids' movie, and they were so pissed! lol.

[Near Lurker]

Asimov also explored this idea in his story "The Bicentennial Man" (source for the movie "Bicentennial Man" starring Robin Williams).  The title character, a robot who had been upgraded so many times that he achieved sentience and looked completely human, and was actually granted citizenship and the same rights as a human being, chooses to "die" rather than continue existing without the human companion with whom he had spent so many years.  I'd forgotten about this story until now.

Is that the film plot or the book plot? As I recall it (from the book, Robin Wiliams eww), he was only granted human citizenship after he'd chosen to die, that decision being what swung the humans concerned with the decision.

"The leaders of nations did chortle,
And scoff at this legal loop-portal,
For beneath all their laws
Lay an unwritten clause...
'No citizen may be immortal.'"

(Yeah, I know, fuckin' Mormons, etc.)

[idontunderstand]

We can't say whether robots can love or not without defining love.

*thinks*

If we define love as the more-than-friends attraction between persons, the probably sexual one, robots can probably not love, since they don't reproduce through sex. However, by that definition I suppose a sentient sex-robot could "love", while a sentient, non-sexual robot could not. In my opinion, this makes it a weak definition.

If we define love as.. somehow friendly feelings towards someone, feeling closeness and trust and dependence towards another person, robots might be able to love, assuming they are sentient beings. The problem is, I figure, that a sentient robot is probably never really dependent upon human beings, seeing as all they really need is a power source. Can someone who doesn't need other persons really love? Now however, love is not the same thing as dependence. It's possible to feel love towards someone without depending on that person. It seems as though the robots in the QC universe are social beings who enjoy other's company, and we can assume they do so by their own choice, not because they are programmed to do so. A being that chooses to live with other beings and interact with them because of their own choice, must have some sort of motive to do so. Love, possibly. Programming to mimic human behaviour, maybe. But Jeph's last twitter post indeed suggests that the robots live among humans because they want to, because they like humans. (on a more realistic note, creating a being with free will is probably impossible.)

The last definition I can think of is the more universal love, the will that drives the universe forwards and makes the plants grow and the sun glow. Not a feeling, but rather the will to live. Every living being has this will, and there is no reason to assume that the QC robots don't.

So I guess.. I dunno? I think they maybe do? 

[HiFranc]

idontunderstand, I remember a discussion about research into love by social scientists on the radio years ago (too many for me to remember the date, year or decade).  Their definition of love was simple:

To understand the needs of another being and to meet them.

Now that I've had time to think about about it I would modify that a bit:

To understand the needs of another being and to meet them (even if it means a cost to the provider).

In this case "cost" could money, could be time, could be heartache, could be anything.  I still feel there's something missing but I can't think of a way of phrasing it that would cover all cases that we would accept as "love".  I think I've got it:

To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest.

I think that that's a reasonable definition of love.  What do you think?

[snubnose]

If there were robots talking and acting like humans, I would think they can love.

This is a natural reaction. Thats why Disney, for example, humanifies animals in their strips. Its what our brain tells us. That animals are somehow human underneath. Even if animals very likely have a different perspective than us, because of their limited mental abilities (compared to us).

Animals are a lot more like people than computers, though. Even the most simple animals can already feel, and they can already hurt.

I'm a programmer, but I cannot program the computer in front of me to feel anything, or to be hurt. It doesnt matter how much memory you give me or how fast you make this computer, it simply wont feel anything, ever.

Mind, I dont oppose that Momo would ACT like she would actually love Marigold. I oppose that there is actual feeling there.

Neurons can fire, not fire, send impulses to other neurons, and change their sensitivity to input. All their activity is some combination of the above. Can machines like us, built from neural networks, love?

http://en.wikipedia.org/wiki/Vitalism#Foundations_of_chemistry. Chemists used to believe there was some magic principle unique to organic molecules that made them different from inorganic molecules, and that they could never be synthesized from non-living ingredients.

I do not believe science has yet understood what consciousness is, and I doubt they ever will. Computers can emulate neuron networks, but that still wont give them the ability to feel, or to hurt.

Computers only have the ability to perform mathematical operations

From a similarly reductionist point of view, human beings only have the ability to perform chemical reactions. How can a collection of chemical reactions love? The existence of sociopathy suggests that, at least to some extent, the ability to love is a learned behaviour, or to put it another way, a matter of programming.

To my knowledge it is more a case of destroyed hardware than the lack of programming. If you disrupt the nerves of a human being or an animal, its possible you can cut off their arm or leg without them feeling anything. Likewise, sociopaths are unable to know consciously what they feel, or to understand other peoples feelings, because of destroyed parts of their brain. They are still able to hurt though.

My issue is simply the claim that was started as early as computers have been known, that somehow making computers faster and more powerful they would turn into something else. Just read or watch 2001 for that one and check out the abilities of HAL 9000. Its more obvious in the book, the movie stays kind of vague about this.

Yet computers did no such thing. They only became faster and better able to store things. They did not turn sentient and show no sign to turn sentient in the near or distant future. Its simply not there. No matter how fast it is, its still just a mathematical calculator.

[HiFranc]

snubnose, you still haven't addressed my criticism of your assertion in that there may be another way of wiring the computer to achieve that end.  You are still thinking that the brains for these robots are constructed in the same way as normal computers are nowadays.  However, if they were constructed in a different fashion (for example by a complex[1] neural network[2]) which didn't rely solely on logic, you might get there.

[1] All complex animals have neural networks.  In the case of a hydra its behaviour is not that far away from a computer.  In our case, we have thoughts, feelings, ideas, inventions, etc.
[2] Electronic neural networks already exist.  I know that the finance industry in Britain uses one (or more) to spot frauds for the reason that neural networks.  Bear in mind that, compared to our brains, is a simple one.

[Skewbrow]

Yet computers did no such thing. They only became faster and better able to store things. They did not turn sentient and show no sign to turn sentient in the near or distant future. It's simply not there. No matter how fast it is, it's still just a mathematical calculator.

Me the mathematician has to disagree. Computers are very good at arithmetic, but mostly useless at mathematics. Granted, I have done a lot of computer-aided mathematical research. But the role of the computer has been to do experiments on my behalf: like test an error correction scheme 100 million times under random conditions, or checking a hunch by exhaustive search in the couple of smallest cases, or.... But a computer won't give me an initial idea to play with, it won't get an insightful "Heureka" moment, or see its way thru a proof, or do anything real math is about. I realize that you equated math with arithmetic the way the general public does, so no harm was done. Your claim just touched a nerve here. So if we agree to emphasize calculator, I won't start a fight.

But life is very complicated (at the molecular level), and so is love. I'm not sure whether it was one of Akima's points, but I think that love cannot be explained in a purely reductionistic way. May be love is learned/programmed? I don't know. Another possibility is that it is, at some level, emergent behavior (a by-product of evolution?). Perhaps the simplest emergent behavior is exemplified by Langton's ant. As a programmer you will probably enjoy the demos! The upshot is that given any (random) initial "universe" an 'ant' (=an automaton much simpler than the Intel x86 processors) walking about in this universe, while following a simple rule, will first spend some time wandering aimlessly, but then eventually will start following a pattern giving the impression that it has a clear purpose and direction.

That example is, of course, anything but convincing. But it does hint at the possibility that out of a chaotic ocean filled with organic molecules apelike creatures capable of love eventually emerge. If love emerges out of combinations of chemical reactions, could it not also emerge out of automata that are somewhat more complicated than that simple-minded ant?

[idontunderstand]

To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest.

I think that that's a reasonable definition of love.  What do you think?

I think you can both be aware of the needs of another being and even meet them, to some extent, without really feeling love towards that being. You could feel responsibility or be liable through your profession, without really feeling anything. So no, I would have to disagree.

[HiFranc]

To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest.

I think that that's a reasonable definition of love.  What do you think?

I think you can both be aware of the needs of another being and even meet them, to some extent, without really feeling love towards that being. You could feel responsibility or be liable through your profession, without really feeling anything. So no, I would have to disagree.

But those two scenarios would be caught by the "genuine caring" clause I added to the end.

{edit} Now I think about it, it may be caught by the "even if it means a cost to the provider" if the person is willing to put their job on the line to get extra help for a person.

[TRVA123]

It depends on the type of definition we are going for. Love is an internal sensation; we might apply it to situations we observe around us, but there is no way of knowing that it is love in the sense we assume. Identifying and defining love is as difficult as defining and identifying depression. People experiences it in their own way, not everyone experiences it, and finding a universal definition for it is bloody difficult and will probably never truly communicate what the emotion is to someone who hasn't felt it.

For identifying love in the actions of others I'd say that this:

"To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest."

is a fair definition, even though it might seem empty to some people, depending on how they have felt love.

[Near Lurker]

Yeah its true you can write a simulation of feelings. But if a human being loves, its not the result of an arithmetic operation. If a human being is hurt physically, the pain they feel is real. A computer wouldnt be stunned or disabled by pain, either.

It's a result of a chemical reaction stimulated by arithmetic operations.  The pain a human or animal feels is no more "real" than that which can be simulated by a computer.

[DSL]

Don't know who said it, but this works for me as a def. of love: When the happiness of another is central to your own.

[rsquared]

I can agree completely with DSL's definition. But this just moves the question sideways: Can robots feel happiness? Snubnose is asserting 'no' (indirectly, perhaps). As we know robots now, I am inclined to agree that they, based on computers, could never truly feel emotions. Maybe somebody someday will invent a new kind of computational engine, a leap over our arithmetic-based PCs the way PCs are a leap over the abacus, that will have that capability.

With sufficiently clever programming, though, I'm sure an arithmetic-based computer could convincingly fake emotions. If an emotion is a state of mind, a complex and nuanced and probably multilayered finite-state machine computer program could evince emotions in a completely convincing way. Which would be "good enough". If you cannot distinguish between "genuine" emotions and programmed emotions, then really, there is no difference. (A Turing test of the heart…)

[idontunderstand]

To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest.

I think that that's a reasonable definition of love.  What do you think?

I think you can both be aware of the needs of another being and even meet them, to some extent, without really feeling love towards that being. You could feel responsibility or be liable through your profession, without really feeling anything. So no, I would have to disagree.

But those two scenarios would be caught by the "genuine caring" clause I added to the end.

{edit} Now I think about it, it may be caught by the "even if it means a cost to the provider" if the person is willing to put their job on the line to get extra help for a person.

Maybe. But I would say you can show care for someone without necessarily loving them. And to say "to love is to care" or "to care is to love", doesn't really make for a definition, since then we would have to define what "care" means in this context.

I like DSL's definition as well, can't find anything wrong with it no matter how I try, hehe.

[questionablecontentfan]

Don't know who said it, but this works for me as a def. of love: When the happiness of another is central to your own.

That sounds like Momo.

[TRVA123]

I went archive diving to find some more "in world" information about the nature of Questionable Content AI,

One where Hanners explains to Winslow where the first AI came from:
http://www.questionablecontent.net/1506

And then several comics where Winslow is talking about various AnthroPC topics:

The negative stereotypes about AnthroPCs: http://questionablecontent.net/view.php?comic=805

What his purpose is: http://questionablecontent.net/view.php?comic=706

Programming and Slavery: http://questionablecontent.net/view.php?comic=1411

[Is it cold in here?]

Even if the difference between a wetware machine and a silicon machine is supernatural, what's to stop God from providing a soul to an entity like Momo who's complex enough to hold one?

[Mr_Rose]

Even if the difference between a wetware machine and a silicon machine is supernatural, what's to stop God from providing a soul to an entity like Momo who's complex enough to hold one?

Well, duh, because that would mean that people aren't a special creation and therefore have no divinely appointed right to do whatever the hell they want to everything else. And since every known* god is a creation of humans, that would never happen.


*If there are alien civilisations out there it would not be that much of a surprise to find that they have or had gods of their own. But we don't know that yet.

[Random Al Yousir]

I always had trouble with the idea of divine intervention.  If she* did a proper job in the first place, there wouldn't be a point in it.  If she has to correct her work now and again, this would suggest lack of skill or information, which violates the whole idea of god.


* The male sex joined life late in the game.  Besides, the bearers of new life are the females.

[pwhodges]

Besides, the bearers of new life are the females.

Not always.  The Seahorse male carries the fertilised eggs in his pouch.

[jwhouk]

Don't know who said it, but this works for me as a def. of love: When the happiness of another is central to your own.

Thinking of others before yourself.

An AI, under that definition, is capable of love - especially if that is how they were programmed. To think, of others, before themselves.

[Random Al Yousir]

Not always.  The Seahorse male carries the fertilised eggs in his pouch.

Huh.  There goes one of my arguments.  However, the other one still holds.

Thinking of others before yourself.

An AI, under that definition, is capable of love - especially if that is how they were programmed. To think, of others, before themselves.

There we are again.  If a being is programmed to care about others, could this really be love?

In my book this should be a decision out of your own free will, but then again ...

Why, when we put our motives to a merciless test, why do we decide to love someone?

[TRVA123]

There we are again.  If a being is programmed to care about others, could this really be love?

In my book this should be a decision out of your own free will, but then again ...

Why, when we put our motives to a merciless test, why do we decide to love someone?

New mothers have a strong chemical reaction when they spend time with their baby. With most of them it triggers a powerful love for the child. Programming?

[Skewbrow]

There we are again.  If a being is programmed to care about others, could this really be love?

They call it the genetic code for a reason?

[Is it cold in here?]

Descartes thought that non-human animals were mere machines and didn't have feelings. I don't subscribe to the belief that humans are unique in having emotions.

[jwhouk]

"The commandments (are all) summed up in this one rule: Love your neighbor as yourself. Love does no harm to its neighbor. Therefore love is the fulfillment of the law."

---

"Love is patient, love is kind. It does not envy, it does not boast, it is not proud. It is not rude, it is not self-seeking, it is not easily angered, it keeps no record of wrongs. Love does not delight in evil but rejoices with the truth. It always protects, always trusts, always hopes, always perseveres. Love never fails...."

"When I was a child, I talked like a child, I thought like a child, I reasoned like a child. When I became a man, I put childish ways behind me."

"And now these three remain: faith, hope and love. But the greatest of these is love."

---

I know - just some tentmaker from Tarsus in the 1st Century. But he had a point.

[Carl-E]

    Let me not to the marriage of true minds
Admit impediments. Love is not love
    Which alters when it alteration finds,
Or bends with the remover to remove:
    O no! it is an ever-fixed mark
That looks on tempests and is never shaken;
    It is the star to every wandering bark,
Whose worth's unknown, although his height be taken.
    Love's not Time's fool, though rosy lips and cheeks
Within his bending sickle's compass come:
    Love alters not with his brief hours and weeks,
But bears it out even to the edge of doom.
    If this be error and upon me proved,
I never writ, nor no man ever loved.

     --Bill Somebodyorother

Ok, it's not a defeinition - just some definite characteristics of love. 

[idontunderstand]

Love is real, real is love,
Love is feeling, feeling love,
Love is wanting to be loved.
Love is touch, touch is love,
Love is reaching, reaching love,
Love is asking to be loved

Love is you,
You and me,
Love is knowing,
We can be

Love is free, free is love,
Love is living, living love,
Love is needing to be loved

[NeverQuiteGoth]

There we are again.  If a being is programmed to care about others, could this really be love?

They call it the genetic code for a reason?

That code creates your brain, not your mind. Your mind isn't your brain; rather you mind is one of the things your brain does. There is absolutely no reason that something other than a brain couldn't be made to do the same thing, just because we don't yet understand anything about the fundamental mathematical structure of a mind.

That said, Random Al is failing here to distinguish between feeling love in a specific instance and the general capacity to love. And in either case, asking if its really love is a bit of a Wrong Question. Love is a multifaceted and complicated construct in our minds, and the only place where it can be real or not is your mind.

Quite simply, if the AI thinks what it feels is real, then it's real. Same goes for all of us, doesn't it?

What matters is not external behavior, but the reason for that behavior. If an AI talks and acts like a human for the same reasons that a human talks and acts like a human, isn't it just as much a person as you are?

Why does it act like it loves? Does it think it loves? Why does it think it loves?
These are the only important questions, and you can't answer them for a human any better than you can for an AnthroPC.

[Arancaytar]

To understand the needs of another being and to meet them (even if it means a cost to the provider) and for that provision to be motivated out of genuine caring rather than narrow self interest.

I think that that's a reasonable definition of love.  What do you think?

I get what you are trying to formulate in that last part, but it appears circular in that "genuine caring" sounds a lot like a synonym for love. Understanding and meeting the needs of a being is something you can observe, as is cost. But when trying to judge what something is motivated by, you have much more difficulty. Finally, if an AI aims to make another person happy because that person being happy and thinking well of the AI will in turn make the AI happy, then some form of self interest is part of the equation, and yet this would be love by most ordinary standards. Otherwise, the only true love would be unrequited love.

Quite simply, if the AI thinks what it feels is real, then it's real. Same goes for all of us, doesn't it?

What matters is not external behavior, but the reason for that behavior. If an AI talks and acts like a human for the same reasons that a human talks and acts like a human, isn't it just as much a person as you are?

Why does it act like it loves? Does it think it loves? Why does it think it loves?
These are the only important questions, and you can't answer them for a human any better than you can for an AnthroPC.

I agree. I would also say that with sufficient complexity, determining what reasons an AI has for its behavior (eg. whether it appears to love because it loves, or whether it appears to love because it is designed to) becomes impossible. Or, to phrase it like Clarke might: Any sufficiently advanced intelligence is indistinguishable from sentience.