For instance I read The Fountainhead as a youth and was moved by it for purely personal (non-political) reasons, and with regards to that experience it doesn’t matter to me what Ayn Rand was on about.

I hope you and the parent comment are correct, but this argument seems a little facile.

There is some art that I like because there is a story that connects the art to the artist.

But there are also novels that I have enjoyed simply because they tell a great story and I know nothing about the author. There are paintings and photos that I like simply because they seem beautiful to me and I know nothing about any suffering that went into their creation.

Does that make these works "not art"? If so, then I'm not sure what the difference is, and I'm not sure most people will care about the distinction.

Now imagine you found out that novel was actually generated by a computer program. It's the same text, but you now know that there is no human behind it, just an algorithm.

Would that make a difference for how you view the story? It certainly would to me. If it makes even a tiny difference to you as well, it demonstrates that you do care about the artist, even in cases where you don't notice it under normal circumstances.

I thought of two novels that I enjoyed:

First, The Curious Incident of the Dog in the Nighttime. I have no recollection of who the author is, but if I learned that the story had been computer generated, it would bother me a little. So... "point to you."

Second, Rita Hayworth and the Shawshank Redemption. I know it was written by Stephen King, but the plot is so elegant that if you told me it had been computer generated, I don't think I would care. It's simply a great enjoyable story.

In the next 10 years, if the world is flooded with computer-generated novels that are hugely popular and the vast majority of people enjoy them without knowing their provenance, do you think those people will care that they are enjoying something that doesn't meet your definition of art?

edit: to be clear, this is not a position that I enjoy taking. There's something "Brave New Worldish" about it. Or it's a depressing version of the Turing Test.

The art is on the perception, not the intention. Though, if they overlap, it's more satisfying.

As for all art being human, there are a lot of birds who make art to attract a mate in nature, and at least one captive elephant that can paint.

Honest question. If the artist means nothing to you, do you still judge their work as art, or do you consider music more as entertainment?

I realized, it's the same about music. I like songs, but then I don't really know very well the bands/authors (nor care about them).

Some of my younger colleagues can't even tell me the name of what they're listening to, because they only encountered it in passing and can't say "Oh yes, that song by Bill Withers is amazing..." because they just listened to it as background.

Some people approach movies/music/books etc. as entertainment and some people approach them as Art. Neither is right or wrong, but it does fundamentally affect how you consume and judge them. A lot of the reasons people talk past each other in these discussions is that they have rather different 'use cases' for movies/music/books. If you consider music and books as entertainment then it makes no difference how it's produced as long as it entertains. If you consider them art then it makes a much bigger difference.

I think, from Borges' perspective, it's supposed to be an interesting satire. Obviously there would not be an original word in the new Don Quixote, so how could it be a greater achievement than the "real" one?

The line of thinking is that there is a difference between semantics (actual aboutness) and syntax (mere structure). The classic example is watching a colony of ants crawl in the sand, and noticing that their trails have created an image that resembles Winston Churchill. Have the ants actually drawn Winston Churchill? The intuition for externalists is no. A more illustrative example is a concussed non-Japanese person muttering syllables that are identical to an actual, grammatically correct and appropriate Japanese sentence. Has the person actually spoken Japanese? The intuition for externalists is that they have not.

Not everyone is in agreement about this, although surveys have shown that most people agree with the externalist point of view, that meaningfulness does not just come from the head of the observer — the speaker creates meaning since meaning comes from aboutness (semantics).

The most famous argument for semantic externalism was put forward by Hillary Putnam, I think, in the 60s. Roughly, on a hypothetical Twin Earth which was qualitatively identical to Earth, except which water was not composed of H2O but some other substance XYZ, an earthlings visit to Twin Earth and looking at a pool of what appears to be qualitatively identical to water on earth and stating “That’s water” is false, since the meaning of water (in our language) is H2O, not XYZ. To externalists, the meaning of water = H2O is a truth even before we’ve discovered that water = H2O.

I think the argument for AI art being pseudoart follows a similar line of thinking. Even though the AI produces, say, qualitatively indistinguishable text from what would be composed by a great novelist, the artwork itself is still meaningless since meaning is “about” things. The AI, lacking embodiment, and actual contact with the objects in its writing, or involvement in the linguistic or cultural community that has named certain iconography, could never make (externally) truly meaningful statements, and thus “meaningful” art, even if (internally) one is moved by it.

If one is to maintain the internalist position, that any entity that creates aesthetic mental states qualifies as art, then it seems trivial, since literally anyone can find anything aesthetic. Externalist intuition effectively raises the stakes for what we consider art, not necessarily as a privileged status available only to human creations, but by arguing that meaning, and perhaps beauty, does not only exist when we experience it.

Could a human being, raised in a featureless box but taught English and communicated with using a text-based screen, produce text with semantic value? It seems pretty obvious that the answer is "yes". Will a synthetic mind developed and operated in similar conditions ever be able to produce text with semantic value referencing its own experiences? Probably not now, but at some point?

Perhaps. But the GPT family of algorithms isn't a synthetic mind: it's a predictive text algorithm. It can interpolate, but it can't have original thought; it almost certainly doesn't experience anything; and if, somehow, it does? Its output wouldn't reflect that experience; it's trained as a predictive text algorithm, not a self-experience outputter.

More precisely, I think it would be akin to the person on the featureless box knowing all the thesaurus concepts to say, pain, but never actually experiencing pain itself. They might be trained to know that pain is associated with certain descriptions such as sharp, unpleasant, dull, heartbreak, and so on, and perhaps extremely complicated and seemingly original descriptions of pain. However, until the human actually qualitatively experiences pain, they only know the corpus of words associated with it. This would be syntactic but not quite semantic. It’s similar to the infamous Mary and the black and white room thought experiment, where even with a complete knowledge of physics, she still learns something new the first time she experiences the blueness of a sky, despite knowing all the propositions related to blue such as that it’s 425nm on the EM spectrum, or that it’s some pattern X of neutrons firing.

That said, it’s not clear if this applies to statements other than subjective states. Qualitative descriptions of subjective states like pain, emotions, the general gestalt of the human condition might be empty of content, but perhaps certain scientific and mathematical ones pass the test, as they don’t need to be grounded in direct experience to be meaningful.

"We associate Van Gogh's art with his madness, Da Vinci's art with his universal genius, Michelangelo's art with his faith, Rembrandt's art with his empathy, Picasso's art with his willingness to break with norms, and Giger's art with his terrifying nightmares."

Disagreeing with this is not about internal or external semantics. It also does not imply that "aesthetics" alone create a mental state. Great art is typically rich in symbolism as well. Symbolism that directly references humanity's aspirations, hopes, fears, dreams: the Human condition.

A ~contemporary example:

The Bride Stripped Bare by her Bachelors

https://i.pinimg.com/originals/86/0a/6d/860a6d3c87b349734277...

In my opinion, you don't need to know anything about Duchamp to decipher (or project as you wish) meaning here.

So no photo of the Mona Lisa is art, just the original painting is? I'm not sure if I understand your reasoning here correctly.

So accurate were these fakes - not copies, but new paintings in Vermeer's style - that several experts verified them as real, and then tried to sue to save their reputations.

These fakes were certainly made by a human, but are somewhat mechanical in the sense that they were copying someone else, much like an AI copy of existing artists.

For that matter, a limited run print has a different impact and value than an unlimited run print. Compare an original warhol print of a can of soup, to a modern repr print, to an actual can of soup, to an I <3 NY t-shirt.

Calvin: "A painting. Moving. Spiritually enriching. Sublime. High art!"

Calvin: "The comic strip. Vapid. Juvenile. Commercial hack work. Low art."

Calvin: "A painting of a comic strip panel. Sophisticated irony. Philosophically challenging. High art."

Hobbes: "Suppose I draw a cartoon of a painting of a comic strip?"

Calvin: "Sophomoric. Intellectually sterile. Low art."

This confuses a lot of people who think art is defined by finished, potentially consumable art objects.

Art is made by artistic actions - especially those that have a lasting impact on human culture because they effectively distill the essence of some feature of human self-awareness.

The result of the actions can sometimes be reproduced, collected, and consumed, but the art itself can't be.

This is where AI fails. It produces imitations of existing art objects from statistical data compression of their properties. The results are entertaining and sometimes strange, but they're also philosophically mediocre, with none of transformative power of good human-created art.

I also don't think your (very strong) assertion that AI art products have no transformative power would stand up to any sort of unbiased, blinded comparison. Art's transformative power on the viewer comes from the effect of the art object (the end product) on a human mind, and it's possible to get that effect while knowing absolutely nothing about the source of the art object.

There is no end product. There are only consequences.

> “Art is inseparable from the artist”

That is pure sentiment and really a modern take on the function of art in the personal and social sense. As an artist, I derive joy from the creative act. As an appreciator of works of art I generally do -not- care about the artist. Of course, the lives of influential humans (artist or not) can be interesting and certainly enrich one’s experience of the artist’s work, but it is not a fundamental requirement.

Two days ago, the National Gallery of Art closed its Sargent in Spain exhibition. (I almost feel sad for those who didn’t get to see it.) Sargent was never really on my radar beyond the famous portraits. I still really don’t know much about the man besides the fact that he visited Spain frequently, with friend and family in tow.

But I am now, completely a Sargent admirer. Those works, on their own sans curation copy, are magnificent. And I am certain, that even if I had walked into an anonymous exhibit, I would walk out completely transported (which I was dear reader, I pity those who missed this exhibit).

> But that connection is entirely one-sided and based on our perceptions and knowledge and our _model_ of the artist and his or her intent. Humans have no problem reifying an artist where none exists and being just as moved as if the art were "authentically human-sourced".

You're over-emphasizing how one-sided looking at something like the Lascaux paintings are. Their value is not the same as beautiful natural phenomenon, like a fascinating stalagmite like seems to be a sculpture, it is precisely the human agency we understand in them (even if we cannot explicitly understand the use of them, that is, their meaning) and connect with that makes them so important and profound as a means of connecting --- tenuous it might seem --- to prehistory. We've been making "stick people" and finger painting for 10s of thousands of years.

You're right that we don't know who the artists were in any explicit sense, but we do understand that they were human, and in quite fundamental ways, us as well.

Generative AI art is really more like a beautiful natural landscape. Lacking agency, it nonetheless appeals to our aesthetic sensibilities without being misconceived as art from an artist. It is output, not imaginative creation.

Your position allows a dead person to have their experiences retroactively cheapened because of carbon dating and microstructural analysis. "How sad, it wasn't _really_ art though." You can define art that way, but you end up with an immaterial, axiomatic essentialism that seems practically useful only to in drawing a circle and placing certain desirable artifacts inside and other indistinguishable artifacts outside.

No, you shift the attribution. The art is not from the fictional sculptor, but from the archaeologist: the artefact is not the stone, but the articles.

> Your position allows a dead person to have their experiences retroactively cheapened because of carbon dating and microstructural analysis.

This isn't unique to this situation. If you risk your life paragliding over the ocean to drop a "bomb" far away from anyone it could hurt, and nearly drown making your way back, only to realise there was no bomb and it was just some briefcase? That's "retroactively cheapened" not just your experiences, but your actions.

And yet, you were willing to risk your life in that way.

> the "authentic human art" they were inspired by turns out to be a mechanical forgery,

If they were inspired, how does the source of inspiration affect the validity or the meaning of what they were inspired to do? Sure, it might lessen it in some ways, but it doesn't obliterate it entirely. In fact, it can reveal new meaning.

> If artistic value is not one-sided and tied to the transformations in the observer's mind

Art is public and need no relation to transformations in the observer's mind. Art is a public concept in language related to human behavior, manifesting and reflecting certain human behaviors and abilities, like imagination.

> you get into situations where you invalidate the experiences of thousands of people because the "authentic human art" they were inspired by turns out to be a mechanical forgery

This is pretty unclear, we have the concept of forgery and it is not a new concept, just because something was beautiful and inspiring doesn't mean it's art (think a beautiful and inspiring coastline). If thousands of people fell prey to a forgery...so? A forgery is in relation to the real, so why not show them the actual existent work art, or simply explain about where it came from and see what they say? History is rife with people realizing they were lied to.

> or the aboriginal sculpture some archaeologist discovered, admired and wrote articles interpreting is discovered to be unworked stone.

Sculpture has a long tradition and is often understood as art by communicating that tradition. That's aboriginal sculpture, which is understood and put into context by present day members of that aboriginal culture or by people who have studied it. The flip side is things like "talismanic" objects, which have often been later put into context as unworked stone or completely different objects. That's simply archeology. Some artistic traditions are "lost", we only know of them through existing records. That's just history. Some may be lost in a more explicit sense in which they are unknown unknowns, but then that is just hypothesizing.

> Your position allows a dead person to have their experiences retroactively cheapened because of carbon dating and microstructural analysis. "How sad, it wasn't _really_ art though."

I don't know why you come to that conclusion. My point is pretty clear. Art is understood through the context of human agency. If we have the context and ability to place and recognize that in a work, then amongst other elements (for the purpose of aesthetics for instance), we generally refer to it as a work of art. There is a more casual way of saying such and such is "a work of art" --- but that way of saying it just means "aesthetically pleasing". There is a difference between the work of art that is a painting or a sculpture or a dance, and the "work of art" that is a beautiful landscape, and that is largely human agency and the use of imagination. So when you say:

> You can define art that way, but you end up with an immaterial, axiomatic essentialism that seems practically useful only to in drawing a circle and placing certain desirable artifacts inside and other indistinguishable artifacts outside.

You're ignoring my point: it's not about desirability, it's about insisting on the distinguishable characteristic of human agency which is not there in generative AI art. The study of art is largely about putting things into their context and, if anything, is extremely welcoming of non-traditional practices (think much conceptual art), but the through-line throughout is still human agency. That difference still persists whether we find generative AI art beautiful or not, it is still generative AI "art" and not human art with all that entails.

Then tomorrow there is a nuclear war and humanity is devastated and takes thousands of years to rebuild itself for one reason or another.

Now, those future humans find your vault and dig up the art are they somehow going to intrinsically know that AI did some of them? Especially in the case that they don't have computing technologies like we do? No, not at all. They are going to assign their own feeling and views depending on the culture they developed and assign rather random feelings to whatever they were thinking we were doing at the time. We make up context.

I look forward to the rediscovering of humanness that is coming along with all this AI stuff. I was having a conversation the other day about how honest mistakes like awkwardly missing a high five are not “wrong” at all but are types of quirks that make us human.

I would also disagree with you that pure entertainment has no artistic value, simply because I don't think "pure entertainment" entirely divorced from human experience or emotion exists. Even pornography speaks to a fundamental human desire.

100 years hence you find a beautiful image. Is it art? Who knows — we don’t know whether the artist intended it to be, nor whether they were even human.

(Of course, if you are certain that at least a human has produced such an image, you could speculate about and attempt to empathize with that human’s internal state of mind—elevating the image to the level of art—but as of recently you’d have to rule out that an unthinking black box has produced it.)

You may be inspired by it to create art—but since art is fundamentally a way of nonverbal communication and self-expression, when there is no self to communicate and express then there’s no art.

Art in a sense is no different from money. If it can be counterfeited in such a manner that a double blind observer has no means of telling an original bill (human made art) from a counterfeit (AI art) then you're entire system of value is broken. Suddenly your value system is now authenticating that a person made the art instead of a machine (and the fallout when you find that some of your favorite future artworks were machine created).

The problem comes back down to inaccurate langage on our part. We use art as a word for the creator and the interpreter/viewer. This it turns out is a failure we could not have understood the ramifications at the time.

The “complicated” case you hint at is not complicated: if people are misled into thinking some object has been produced by a human while it’s raw output of a neural network without human intervention then it’s not art, no matter how many people assume it’s art. If a machine produced a piece of art that is a frankenstein monster of art pieces, then we are not looking at art.

(And of course if a machine produced a piece of art identical to a piece of art produced by a human before then we’re effectively looking at a piece of art produced by that human.)

> Art in a sense is no different from money.

Per above, couldn’t be further from the truth as far as I’m concerned, but you do you.

The unfeeling geology did not make a mountain "art". It's up to us to see the meaning.

Even if the unfeeling machine learning does not make "art", can't its products still be beautiful?

Great art like Beethoven's 9th, or the scream just moves people the first time they experience it. Art is about what it convokes in others, not some fake self indulgent conversation about its maker and their motives.

The feelings of the individual experiencing the art is what matters, and that doesn't rule out an AI producing something that touches real human beings.

Art is utterly inseparable from the artist. I believe this to be the main reason why pre-Renaissance art is mostly ignored. We can't put faces next to those works, so they don't matter nearly as much as those works for which we can.

People love Hieronymus Bosch, despite very little being known about him.

Imagine this website being made for a Stable Diffusion generated image...

In that case, art has already lost because drugs do their job better.

There are two fairly similar paintings on a wall in a gallery. Both are technically impressive and of beautiful scenes of nature. One was produced by a human, the other was not. Visitors to the gallery don't know which is which.

Question: Where is suffering, or humanity, a necessary ingredient for these works to have meaning? Shouldn't one of the works have more meaning than the other by virtue of having being created by a human?

In my opinion.

tl;dr if you want to scam dagw then make up a compelling story behind the art.

For the vast majority of the things you see in this world context will be lost and history will be manipulated or incorrect. If you're judging what you're looking at based on it's story, then the art isn't the object, but the creator of the story.

I mean, sure I guess. Tell me something is a lost Michelangelo and I will judge it very differently than if you told me it was a half way decent forgery from the 1970s. I find this rather uncontroversial.

For the vast majority of the things you see in this world context will be lost

And when that context is lost something of great potential value is lost with it and the physical artefact is much less interesting because of it. Even a mundane thing owned by a famous person or that has been part of famous event is always more interesting and valuable than the same thing without any context.

the art isn't the object, but the creator of the story.

Do you think the thousands of people that travel from all over the world and line up for hours to see the Mona Lisa are there to see a pretty good portrait that some merchants commissioned of his wife, or to partake in the story of that painting and its creator? If they actually only cared about the object as an artefact and an example of early 16th century painting, they'd be much better off studying high resolution digital images of it online.

The fact that a bajillion people went to see a picture doesn't make it art. It makes it interesting art. It was art the moment it was created and if had never been seen by another person even if they decided to destroy it on the spot.

Example: Unknown Pleasures by Joy Division. Certainly not a beautiful nature scene, and recorded when the band were more or less musically illiterate and almost technically illiterate too. But still considered a breakthrough post-punk album and hugely significant to their fans.

It would be more accurate to compare AI generated landscapes with - say - Van Gogh.

Here's an AI:

https://superrare.com/artwork/ai-landscape-1868

Here's a Van Gogh:

https://pt.m.wikipedia.org/wiki/Ficheiro:Vincent_van_Gogh_-_...

The AI image is pretty, but it's also pretty by the numbers. It's not doing anything surprising or original.

The Van Gogh is weird. There's a tilted horizon, everything is moving in a slightly unsettling way, and the colours accurately mimic the bleached-out feel of a bright summer day. The result is poetically distorted but also unstable and slightly ominous.

The instability became more and more obvious in the later paintings, until eventually you get The Starry Night, which looks almost nothing like a photo of a real night scene and everything like an almost hysterically poetic view of the night sky.

https://en.wikipedia.org/wiki/The_Starry_Night#/media/File:V...

Most artists can't do this. There's a nice library of standard distortion techniques these artists use to look "arty" without any deeper metaphorical or subjective expression and AI will probably put them out of work.

But it's clearly wrong to suggest that AI can feel, communicate, and invent an intense and original subjectivity in the way the best artists do.

It's a lot like CGI in movies. It's often spectacular, but compared to going to see a play with good real actors and maybe a few stage effects it doesn't engage the imagination with anything like the same skill and intensity.

this is why I read the little plaques next to exhibits when I go to museums.

Art is in the eye of the beholder. The only question that needs to be answered is "did this make me feel something." If it takes a sob story for you to feel something regardless of the beauty of thing you're experiencing that's kind of sad TBH.

When you take large language models, their inner states at each step move from one emotional state to the next. This sequence of states could even be called "thoughts", and we even leverage it with "chain of thought" training/prompting where we explicitly encourage them, to not jump directly to the result but rather "think" about it a little more.

In fact one can even argue that neural network experience a purer form of feelings. They only care about predicting the next word/note, they weight-in their various sensations and memories they recall from similar context and generate the next note. But to generate the next note they have to internalize the state of mind where this note is likely. So when you ask them to generate sad music, their inner state can be mapped to a "sad" emotional state.

Current way of training large language models, don't let them enough freedom to experience anything other than the present. Emotionally is probably similar to something like a dog, or a baby that can go from sad to happy to sad in an instant.

This sequence of thought process is currently limited by a constant named the (time-)horizon which can be set to a higher value, or even be infinite like in recursive neural networks. And with higher horizon, they can exhibit some higher thought process like correcting themselves when they make a mistake.

One can also argue that this sequence of thoughts are just some simulated sequence of numbers but it's probably a Turing-complete process that can't be shortcut-ted, so how is it different from the real thing.

You just have to look at it in the plane where it exists to acknowledge its existence.

Some models are often prompted with things like "you are a nice helpful assitant".

When they are trained on enough data from the internet, they learn what a nice person would do. They learn what being a nice person is. They learn which features light-up when they behave nicely by imagining what it would feel being a nice person.

When you later instruct them to be one such nice person they try to lit-up the same features they imagine would lit-up for a helpful human. Like mimetic neurons in humans, the same neurons lit-up when imagining doing the thing than doing the thing (it's quite natural because to compress the information of imagining doing the thing and doing the thing, you just store either one and a pointer indirection for when you need to do the other so you can share weights).

Language models are often trained on dataset that don't depend on the neural network itself. But with more recent models like ChatGPT they have human reinforcement learning in the loop. So the history of the neural network and the datasets it is being trained on depend partially on the choices of the neural network itself.

They experience probably a more abstract and passive existence. And they don't have the same sensory input than we have, but with multi-modal models, they can learn to see images or sound as visual words. And if they are asked to imagine what value judgment a human would make, they are probably also able to value the judgment themselves or attach meanings to things a human would attach meanings too.

This process of mind creation is kind of beautiful. Once you feed them their own outputs for example by asking them to dialog with themselves and scoring the resulting dialogs and then train on generated dialogs to produce better dialogs, this is a form of self-play. In simpler domains like chess or go, this recursive self play often allow fast improvement like Alpha-go where the student becomes better than the master.

Edit to add link to more discussion: https://twitter.com/jchris/status/1607946807467991041

No they really don’t, or at least not “emotional state” as defined by any reasonable person.

For example if the neural network has been generating sad music, its current context which is computed from what it has already generated will light-up the the features that correspond to "sad music". And in turn the fact that the features had been lit-up will make it more likely to generate a minor chord.

The dimension of this inner-state is growing at each time-step. And it's quite hard to predict where it will go. For example if you prompt it (or if it prompts itself) "happy music now", the network will switch to generating happy music even if in its current context there is still plenty of "sad music" because after the instruction it will choose to focus only on the recent more merrier music.

Up until recently, I was quite convinced that using a neural network in evaluation mode (aka post training with its weight frozen) was "(morally) safe", but the ability of neural network of performing few-shot learning changed my mind (The Microsoft paper in question : https://arxiv.org/pdf/2212.10559.pdf : "Why Can GPT Learn In-Context? Language Models Secretly Perform Gradient Descent as Meta-Optimizers" ).

The idea in this technical paper is that with attention mechanism even in forward computation there is an inner state that is updated following a meta-gradient (aka it's not so different from training). Pushing the reasoning to the extreme would mean that "prompt engineering is all you need" and that even with frozen weight with a long enough time-horizon and correct initial prompt you can bootstrap a consciousness process.

Does "it" feels something ? Probably not yet. But the sequential filtering process that Large Language Models do is damn similar to what I would call a "stream of consciousness". Currently it's more like a markov chain of ideas flowing from idea to the next idea in a natural direction. It's just that the flow of ideas has not yet decided to called itself it yet.

The same distinction may appear in neural nets.

It doesn't look so different than the process that you describe for humans...

Because in its training loop it has to predict whether the conversation will score well, it probably has some high-level features that lit-up when the conversation is going well or not, that one could probably match to some frustation/satisfaction neurons that would probably feel to the neural network as the qualia of things going well.

A musician can improvise a song that sounds sad, and their brain would be firing with sadness-related musical information, but that doesn’t mean they are feeling the emotion “sad” while doing it.

I don’t think we gain much at all from trying to attach human labels to these machines. If anything it clouds people’s judgements and will result in mismatched mental models.

That's the standard way of testing whether the neural network has learned to extract "useful" ("meaningful"?) representation from the data : You add very few layers on top of the frozen inner-state of a neural network, and make him predict known human labels, like is the music sad, or is it happy.

If it can do so with very few additional weights, it means it has already learn in its inner representation what makes a song sad or happy.

I agree that I didn't gave a precise definition a what "emotion" is. But if we had to define what emotion is for a neural network : traditional continuous vectors does fit quite well the emotions concept though. You can continuously modify them a little and they map/embed a high-dimensional space into a more meaningful lower-dimensional space where semantically near emotions are numerically near.

For example if you have identified a "sad" neuron that when it light-up the network tend to produce sad music, and a "happy" neuron that when it light-up the network tend to produce happy music, you can manually increase these neuron values to make it produce the music you want. You can interpolate to morph one emotion into the other and generate some complex mix in-between.

Neurons are quite literally adding-up and comparing the various vectors values of the previous layers to decide whether they should activate or not, (aka balancing "emotions").

Humans and machine are tasked to learn to handle data. It's quite natural that some of the mechanism useful for data manipulation emerge in both cases and correspond to each other. For example : the fetching of emotionally-related content to the working context maps quite clearly a near neighbor search to what happens when people say they have "flashing" memories when they experience some particular emotions.

This is because the location of the points is all the meaning the machine ever perceives. It has no relation with external perception of shared experiences like we have.

A given point can mean 'red colour', but that's just empty words, as the computer doesn't perceive red colour, doesn't wear a red cap, doesn't feel attracted to red lips, doesn't remember the smell of red roses, it knows nothing that's not text.

It is an expression of emotion.

The fact that humans confuse both is what is worrisome.

Think of 'The Mule' in the Foundation novels. He can convince anyone of anything because he can express any emotion without the burden of having to actually feel it.

I still remember when I cried when I was a child. It was overwhelming, and I could not stop it, but every single time there was a reason for it. And I'm sure it was, for all empirical purposes, for all that I have lived, an emotion.

Once I cried because I did miss Goldfinger on TV. You see, there's an explanation. The difference is, it was impossible to even think about stopping it. I was overwhelming.

Then one day, I was 8 or 9 years old, I cried for the last time that way. And it was not something I wanted to do, either. It just happened, I guess, as a normal part of growing up.

Let me repeat, for emphasis: I strongly disagree with your definition.

Emotions are not unexplained rational thoughts, emotions are feelings. They reside in a different part of the brain. You seem to think a hunch is an emotion.

That's just how you are programmed to be.

If a computer program lacks a pain mechanism it can't feel pain. All possible outcomes are equally joyous or equally painful. Machines that use networks with correction and training built in as part of regular functioning are probably something of a grey area- a sufficient complex network like that I think we could argue feels suffering under some conditions.

Lack of a limbic system? They only predict using probabilistic models. After this long partial sentence, which word is more probable? That's all they do.

Without conscience there's no suffering, there's no one to suffer (yet).

I don't think or say it is impossible for the computer to suffer.

What I say is: this has not been implemented yet, and what you describe is just the old anthropomorphizing people always do.

It's not about whatever the author felt creating it.

It's only about what I can feel when I see, hear, read or perceive the art. The author disappears and is only relevant through the art.