Well, im still amazed at human brain. In several ways. One thing that amaze me is the development of skills over the years. How that child increase his/herd ability to produce art in original ways. Now that we read every day of the huge databases that are need as input for ai to be able to produce something, i amaze how humans can produce thing from nearly nothing. Really, there is a mistery involved in our brain, our beings.

You seem to be asking the question until you get the answer you want, William. You have not responded to some very detailed responses to your question, possibly because you don't like them, or maybe because you cannot find a way to counter them. You had to be prompted to even acknowledge them (see your response in post #26). I find this quite rude.

Nothing that CoPilot produces is art. Please take a moment to read and respond properly to the very detailed responses above: people have given generously of their time to write them, and it feels as if you are choosing to ignore them.

Your threads do tend to be very one-sided: all you are ever really interested in is people agreeing with you.

I don't get it, William. You asked a question and people have answered it - wouldn't you expect an acknowledgement if it were the other way around?

The comments about photography are equally relevant to any art.

In post #28 you have effectively asked the same question again: are you dissatisfied with the responses so far? Because it seems to me that your question has been answered and most people seem to agree that AI-generated artwork is not art.

You seem to be treating us as if we were AI bots: you seem to be asking the same question in a different way hoping to get a more favourable (to you) outcome. This doesn't therefore feel much like a discussion (since you aren't actually joining in with the discussion).

No one should expect a response to each and every post, but many of the replies to your posts have included direct questions that you seem to have ignored. Perhaps they were awkward questions, but I know from past experience that when you’ve been asked a question which you’d rather not answer you’re perfectly capable of saying so explicitly.

I have looked through the posts again.

Where are any questions asked?

I have read through the posts again and, apart from questions that I have asked, I have found only one question.

That question was asked by Alfred in post 24 and I answered in post 25.

My apologies, William! I got it the wrong way around. You haven’t been ignoring questions that others have asked you, you’ve been asking questions and then — as Ali has pointed out — failing to acknowledge the answers.

I am grateful that people have replied and are participating in this ongoing discussion thread. For the avoidance of doubt, I am not dissatisfied with the responses. I am a researcher, I am interested to know what opinions people have. I am not trying to make a pitch for a particular opinion to become accepted.

The question is asked as a lead in to a supplementary question.

Is the answer the same if a human artist had produced the same picture?

I ask that because the human artist would have trained using the same, or similar, information that was used to train Copilot.

So, if a human artist, as part of his or her training learns to paint in the same style as another painter, is that result art if the picture is an original composition?

Or is it not art on the basis that art needs a new concept?

Before directly answering that, let me go deep into certain matters. There are several huge differences between what generative AI outputs and the work of a painter. Most AI defenders want to ignore these facts, somehow. Some in bad faith, but others just don't know better because they have not studied and trained art skills for decades, so it is understandable.

Gen AI cannot draw nor paint, full stop. Does not really know about perspective, does not understand anatomy or color theory, and...does not feel nor can think.

So, the "information" needed to process is not the same by a huge margin, in kind and quantity. Plus, painting is far from being solely processing information: that is the very basic thing gen AI does, while painting is a lot more involved (and better). An artist drawing with a pencil or painting with a brush and oil paints is not simply good at it because this person has "processed information." Additionally, the argument of "but artists copied from masters" is very false, too. I learned from reality, not from other people's art, and from understanding anatomy, etc. My art does not come from "scraping" any artwork from others. 

An artist trained in all those skills during years: anatomy drawing, perspective, composition, color, lighting, and very importantly, created a particular style, own expression, which serves to express one's feelings and ideas: something the AI algorithm does not even have. As I have explained in several instances before, this strongly distinguishes art from an automated process.

The AI does not "learn" all those skills. Indeed, if it did, you would have never seen a hand with 7 fingers in the AI outputs (to mention a simple example, but a lot more mistakes happen). It does not produce so many errors now, but not because it now "understands" anatomy or ANY of those matters. It understands nothing. They (the developers) just fine-tuned it to try to not output that error so often, but the AI does not really know why.

So...there you have it: generative AI does not use at all the same methods as an artist. It is really a glorified mosaic maker of blurry pixels that is helped by an immense database of unlawful origin; it just predicts the next pixel based on probabilities extracted from all the human work that it has in that humongous database. It's making a puzzle by mere imitation (helped by randomness), without understanding a heck about the puzzle.

But also, it does not "feel" nor "think." This is another so important difference that I don't know why there's even debate about it. Obviously, this all comes from an attempt to simply disregard definitions of art from all that has been academically established about the matter, just for mere interest. But the facts, and what things really are, cannot be twisted.

You seem to hyperfocus on the "result." But this is the output of a system that plagiarizes artwork in a well-hidden way (and which used a legal gray area—for now, but this is starting to be regulated—to scrape the data). But art is never ONLY the result. Art is the entire process (by definition) and a language to express emotions. AI has no emotions. When someone prompts something to Copilot or etc., the individual is not using a brush nor intervening in the art process at all. The "prompter" is only doing the work of a client when ordering/commissioning a piece of art. I explained this in detail earlier, though.

So to answer specifically this supplementary question, I needed to preface it with the above paragraphs.

If with this humans-only example, you want to reach the conclusion that AI is like a human artist, as after all it is using the same information (other artwork from many artists, the masters, etc.), I answer to you that no, it is not using the same information. And more importantly, not in the same way. Firstly, a human painter cannot "scrape" that much data, and especially key, the human artist does NOT need it. Does not need gazillions of data to predict what pixel comes next, because a painter works at a MUCH higher intellectual level. That is, understanding anatomy, perspective, composition, color, light, etc., the "human artist" can imagine and paint from scratch. AI can't do this.  But also, to reach that output, AI is not doing any method comparable with art, by any stretch, as I have methodically explained.

If we were to take your latest question in a literal sense, though... An artist producing the exact same painting as another from another painter is mostly a copyist or a plagiarist. Both (the first, doing it legally, with permission, and the second, not at all) require a LOT of artistic skill, but they are not expressing their own human experience and emotions or ideas; they are merely copying someone else's. Which does not match the definition of art creation. Does it have merit? A LOT more than someone using AI, that is for sure, as to make a good copy, you need to first be a really good painter, great in every of the mentioned matters. And if it is a copy of the great master painters from other centuries, it typically requires mad skill.

Plagiarism is not considered art per se, though. And also, it often allows the original author to take legal actions (which ends up favoring the original author practically always), and that is one of the several reasons why what AI does with the artwork it scrapes should be prosecuted legally.  Obviously not the users of a service or app (i.e., you), but the creators of that product and the companies having done the scraping without permission from the painters and other copyright owners should answer for their actions.

Side note: Gen AI cannot produce an exact image from one prompt to another, as it's based on a random seed. Doesn't add much to this conversation, only to mention that it really can't produce two images exactly the same twice. A human copyist could do a much better job at it (and there are amazing professionals doing it). So, the "thing" is a bad copyist, too. 

An artist (or a "human artist," as AI fans like to say, as if there was another possibility...) is a lot more autonomous than the automated algorithm of an AI, because, well, the "human artist" can actually draw and paint and has all the skills and knowledge that I have referred to. AI is neither "imagining things." That is not really what is happening.

I hope it answered the new question.

Here is a picture derived today from a PDF docutment that I produced some time ago using probably Affinity Designer or possibly Affinity Publisher, I think I used various shape tools and the Pen Tool. No AI involved.

Is it art?

Good morning, Ali.

Thank you for posting.

Indeed.

So the test of whether some artwork is art is not a matter of its quality.

Reaiised to the extent of my ability to do so at the time.

Conveying the general layout of what I had in mind.

Thank you for posting.

After my post and before seeing your second post of today, I have been doing some experimentation using Copilot.

I hope to post transcripts and images.