Could machines feel joy?

Nice summary of different positions.

I quite like the idea of quantum consciousness, though it does tend to just remove the idea into the realm of the un-understandable, conceptually that is - which means that it is basically just being used by some as a modern, more palatable word, for magic. It still seems close to the ideas in neural networks, but with a quantum layer added overtop to include a few of quantum mechanic's funnier aspects. One thing i think is interesting though is that it would materially explain many apsects of spirituality, but not necessarily in the way people might like. It would allow for many funkier things to be explainable, but would also render them as another aspect of brain physiology, albeit it quantum brain physiology, whereas more magical ideas such as soul imbuement might be nicer as they would be less linked to this universes physics.

The article seems best to me at showing the lack of clear terminology between different groups. Consciousness has been such a historically prized aspect of our existence that there have been so many claims laid to it that it is no surprise that this is what we all face. Perhaps it would be good to start by agreeing what we even mean by consciousness.

The computer in front of you has no idea that it is a computer, one that uses electronics and electricity for power. It also has no idea that there are other computers that rely on neurons and on chemistry for their power supply.

Further, the computer in front of you has no desires, and no sense of the desires of others.

My guess is that once these components are added to an AI program, then you will have a conscious computer. I think this could have been done a long time ago, perhaps even with lisp on AT class machines, but no one thought of doing it.

The report from this conference shows that no one has thought of trying it yet.

Joys are defined in terms of desires.

Another item: the computer in front of you has no concept of existence. Perhaps that would be statement one: existence exists. There are some more initial statements, but they are too obvious to set out here.

One interesting possibility is that we might create human-level, humanlike AI systems before we puzzle out the mysteries of consciousness. These AIs might puzzle over their own consciousness, much as we do over ours.

For some re4ason this reminds me of something transmitted by the novel Memnoch, by Anne Rice. Memnoch is part of the vampire chronicles Rice is so famous for —you know, Interview with the Vampire & all that— and it was very appealling to me because it pondered with the purpose of creation.

Memnoch —which is the name Rice gives to Satan— ponders on whether God initiated the process of material creation to discover where he came from. So, getting back to consciousness in AI, it might be that the only way we'll know what consciousness really is, it's when we're able to create a synthetic version of it!

Now, whether AI might spawn emotions, I might argue that an emotional response it's an evolutionary necessity in order to 'program' an organism to seek its self-perpetuity. We have come to believe that an AI's version of intellience would be something very similar to the Vulcan ideal: cold logic devoid of emotion; but it might very well be that Natural selection would disfavor such an intellectual approach.

Remember the movie "The Devil's Advocate" with Al Pacino & The One?: "Vanity is definitely my favorite sin. it's so basic: self-love". Maybe in order for an intelligence to survive, a bare minimum of vanity is in order... The 3 Robotic rules of Assimov seem to imply this, as they seeked to program robots to try to preserve their physical integrity, although they created an "unnatural" buffer of giving precedence to the robot's makers needs over the robot's itself.

maybe religion is our own human version of Assimov's 3 rules :-P

How could you create a synthetic version of something you didn't understand in the first place??? I don't see how this would be possible.

Regards, Kathrinn

I know this article I'm linking below is about AI rather than about feelings but it does put into perspective some of the more "optimistic" claims for programmed intelligence.

http://www.salon.com/tech/feature/2003/0...

Personally, I'm more sympathetic to the views of the writer of the article in the original post than I am to Dennett's views. He (Ben Goertzel) calls himself a panpsychist whereas I tend to use the term idealist to describe myself. The difference in definition is subtle and probably summed up nicely in Wikipedia:

My idea of mind is that it is universal: that all that exists, exists within mind and IS mind.

Dave.

for this to happen, then we need to proove feeling as a solely biological function only. Then we would have to imitate this function through electronics. I believe that the biological effects of feeling are rather an effect not a cause. So what is driving the cause to produce these effects? That is the question. This unseen, unknown driver can never be duplicated by circuitry
, diodes or otherwise.

Clearly, we are not even asking the question clearly:

Do we want to make a machine that feels? Or do we want to make a machine that acts like it feels?

The second one is a lot easier, we just fake the symptoms. That's been done with paintings. The Mona Lisa looks like she is thinking about something, but you can't persuade me that the canvas and the paint have any emotions.

If we want to do the hard one, actually make artificial emotions, that still leaves a lot of open questions.

Suppose we can do that, make a machine feel. How would we know if that mechanism is anything like the reasons we feel things?

However even if it is not, especially if it is not the same, we will learn a lot about ourselves.

Who knows!

Some people apparently don't feel emotions, or lack certain ones. Dysfunction might be a clue, a little crack to spread a little light on our own emotions.

I do wonder what aliens might make of our emotions. What if they had hundreds more and looked on ours as some sort of primitive display of limited abilities, like we might look on a dogs.

Also, are we so confident of our own emotional abilities. If we are to somehow, in principle, disembody the mind from the body then which bits are the mind and which bits are the body.

In terms of the hodge podge of evolution the body came along well before the complex neural circuitry we call our brain, or even the primate brain, or even the mammalian brain. As all this developed the brain was never given supremacy over the body as evolution works on all parts at once. Instead we have a symbiosis between the whole organism and the bodies different parts, which have different requirements, developed the means to affect the nervous system.

So we 'feel' hunger and it grows to consume our thoughts, or we 'feel' our reproductive hormones spike and they give us 'desires'.

This has all evolved alongside each other so that it becomes hard to look at feelings/emotions as just the mind. It is the whole organism.

This view of the brain as a neural network with all thoughts and feeling confined to the network is ok as it goes, but is a little limited here. All those chemicals, hormones, acting on the brain complicate this.

As for programming this, i really don't know. I have been programming since i was seven and wouldn't know where to begin. A programming language is much like a human language. Part of me wants to ask whether we would expect to be able to write emotion in english and expect it to be any better than in Basic or C++. The command set of a computer language really just describes logical rules for the computer to follow, add this, subtract this, copy this to this memory address. Each command is a set of rules like this.

This top down approach doesnt seem to be a good way to me, my intuition says that unless you write a new instruction set based on different principles (something biological?) then you are limited to mathematical rules and copying memory addresses.

The bottom up approach is much more interesting. Simple neural networks are already able to move around 3D environments and find 'food' (or electricity). These 'neural' networks are deceptively simple, but naturally solve problems. A computer program could do this too, but the point is that there is no program.

As for awareness, consciousness and experience. These are very subjective terms. Very often it seems like people are tempted to place us up near the top of some sort of consciousness ranking. I know that isnt always the case here, though it is sort of implied in ideas that we 'havn't woken up yet spiritually'. The implication is that humans, in our makeup, have the ability to actually wake up and be 'more conscious/aware'. Or that we did it in the past and don't do it now, rather than simply being human no matter what we do.

It definitely feels like i am self aware. I might say that a dog is less aware than i am, but perhaps it's experience is just different. Perhaps it is self aware, but just in a doggy way rather than a human way. Our tests of self awareness, such as recognising oneself in the mirror, are very human-centric. I wonder if a dog can be self aware without all the human traits of human self awareness, or simply without the brain function to recognise oneself. Can we be selfware, but just internally and on our own? If a dog thinks of itself as 'I' when it wants a biscuit then would that be enough? Besides, my 6 month old son hasn't really jumped at his own reflection in the same way he will in a few of months, yet i am not tempted to say he is not self aware even if he might not be. I have some sympathy for the idea that we learn the 'I' as we go along.

Some scientists are about to replicate the brain in a supercomputer, all the trillions of neurons. Maybe that will spread some light, but i would have thought you would need to understand the chemistry and overlay this on the neural network before any real understanding could be found on how it might work. I struggle to think how we might even phrase the question without understanding better what we even mean by these concepts.

Well hell, to me this begs the question, just how common an emotion is 'joy', anyway?

There's happiness, contentment, etc. But "joy" is a pretty strong word, and I think if it was at all common, we'd have a much different society.

It barely exists among humans in the first place.