The problems with AI art

[Oniya]

So, a couple of places that I frequent have been having this discussion, and I wanted to bring it up here.  There are a number of web-programs that will generate art/faces/etc, based on user prompts and things they have 'learned from'.

There are a couple of potential issues with this.  One that is a little scary is that if you put art/pictures on the Internet, these web programs can appropriate them as 'learning material'  One generator, called Lensa, has been called out for 'scraping' copyrighted images

From that article:

Adding a watermark may not be enough to protect artists — in a recent Twitter thread, graphic designer Lauryn Ipsum listed examples of the “mangled remains” of artists’ signatures in Lensa AI portraits.

One artist, trying to check to see if her own works had been scraped, found pictures from her own medical records in one of the 'training data sets'.  From an end-user point of view, there's no way of knowing what images the generator uses to compile an end-piece, and whether the subjects of those images knows or would want their image included. 

I don't know about folks here, but that's kind of creepy to me.

[RedRose]

This is worrying. How does medical data end up like this? I mean, I've heard interesting things from docs about patients - but this seems one step worse.

[Oniya]

The article says that the woman's doctor had died sometime between when the picture had been taken, and when she found it in the dataset - there's no way to definitively stamp when the image made it onto the 'net, but she presumes that it was after his death.  (Mind you, when I worked IT at a lab, we zero-formatted any computers that were retired prior to disposal or donation.)

[Callie Del Noire]

Dead or not someone is guilty of a HIPA violation.

[Azy]

I'm on the older edge of the Millennial generation and I should be all pro technology.  The average person my age is.  I admit that technology is useful in ways, but this is one good example of why I don't like  everything turning digital.  It's so easy for private information to be stolen.  I just had my credit card information stolen a few days ago.   

[RederReds]

Oh boy, as an artist I have opinions on AI art. The negatives far outweigh the positives, in my mind, but I think that ultimately the technology is neat but absolutely misused. It's good for quick ideation of ideas for artists and non-artists alike but... that's about it. That and it's not really AI, it's machine learning. Many people hear AI and believe that the program can think on its own when it's really not. Machine learning works on labeled datasets which have to have hundreds, if not thousands, of images to function correctly.

So, for example, you create a dataset with a bunch of pictures of college football games with them labeled based on the college. From there the program does any number of different math computations (convolution, mainly, but dear god I do not understand it despite taking a class on ML) and generates, essentially, an equation of what the optimal 'values' for any given feature on an image are. This allows the algorithm to now look at an image and identify key features such as color, shape, textures, edges, etc as it has an idea of the average 'values' of a given image. Now, art generation isn't the exact same, but the general process is. You get a large dataset of labeled images, run them through some sort of machine learning algorithm, and then you now have an algorithm that has an idea of what an 'apple' looks like. The only problem is you need a LOT of data to get anything near good so... Thus come the issues.

The datasets that these algorithms use are gigantic and sourced from scraping the internet for art without artists consent and without care for what it grabs. This, of course, has many artists unhappy because it's essentially mass larceny. There is no way to protect your art if you post it anywhere and even generators who are trying to be more ethical only have an "opt-out" feature where artists have to go piece by piece and tell the programmers to not use their work - talk about draining. Another key part of this is artists aren't being compensated in any way. Their art is being used to train an algorithm which is then being sold to others for use or, worse, being used by others and then re-sold as "original art." People have also done studies on the datasets used to train some of the most widely used generators and it's problematic to say the least. (Twitter thread link) Like with a lot of technology these algorithms often show the prejudice of society in them including sexism, racism, etc. For example if you search "flight attendant" in the dataset you are presented with women who are often sexualized or outright pornographic. Meanwhile if you search for a common male name you get no such results.

This doesn't even cover the general cultural attitude problem with the vocal subsection of the internet that tends to utilize these generators. I want to preface this by saying not everyone who uses an AI art generator is this way, but, there is a vocal subsection who are. There have been art models trained on the recently passed Kim Jung Gi hours after it was announced that he had passed on, models trained on multiple prolific living artist's work against their very vocal and very public will to not have their art be a part of these models, and models trained out of spite on the previously mentioned artists' work as a way to 'put them in their place' and impose 'supremacy' over them. Hopefully it can be seen why these decisions are not only anti-artist but also super disrespectful to the many affected. Beyond that, the general attitude these people show to artists speaking out about their work being stolen ranges from misinformed to outright threatening and toxic. I've seen people tell artists that they deserve to have their art stolen because they're "gatekeeping" art (?????) as well as people outright tell artists that they will be homeless and out of a job once the "AI revolution" comes. Now, some of this is general internet culture, but in my opinion it doesn't excuse the behavior.

One thing I will note, however, is that AI art tends to only be an issue for a few reasons:
1) The training data is taken from artists without consent or reimbursement
2) People are using AI art to phase out paid roles for artists
3) No one single person has the time to verify training data so you get, as mentioned, medical photos in the dataset as well as general prejudice being proliferated

[Kuroneko]

As another artist, I agree with RederReds completely.

I love technology, and I use digital tools, programs, etc to make artwork, concept art, and costume renderings, but the outright theft of images without the consent of artists is theft, plain and simple. It's unethical. The argument that all 'all artists learn from the work of others' presented by many AI defenders is also disingenuous. Do artists learn by doing studies of existing art? Absolutely. I have my own students do it to learn techniques and important art concepts. Does that mean I can say it's my own work? No. It's a study. It's a learning tool. It's practice, and artists spend years training to develop their own style. If it's theft for a human artist to copy another artist's work, then it's also theft for AI programs to do it. Many artists have shown that they've lost jobs because these programs harvested their work and can now replicate their signature style.

The best, most ethical use of this technology I've seen is using your own image as a base to quickly generate variations. I can see how this would be useful for a concept artist to speed up their workflow. But until the ethical questions regarding the image scrapping are answered, it's too problematic. Once you know these images were created using stolen images, they're very hard to appreciate, no matter how beautiful.

The fact that images of medical records were part of the Stable Diffusion image learning set is frightening.

[Vekseid]

The genie is not going back in the bottle. People need to be aware of what it means and what sorts of discussions need to be had, given the new reality.

To me, AI is a force multiplier. For artists, the end result I see is they will make images and use AI to generate thousands of permutations of each of them. AI art is still very much in its infancy, and 'prompt engineering' can only get you so far with it.

Oh boy, as an artist I have opinions on AI art. The negatives far outweigh the positives, in my mind, but I think that ultimately the technology is neat but absolutely misused. It's good for quick ideation of ideas for artists and non-artists alike but... that's about it. That and it's not really AI, it's machine learning. Many people hear AI and believe that the program can think on its own when it's really not. Machine learning works on labeled datasets which have to have hundreds, if not thousands, of images to function correctly.

This sort of objection has been made to every advancement in artificial intelligence over the past sixty years. The first artificial neural networks couldn't even perform a XOR operation. Now I can tell one to show me a picture of a chitinous apple being fired out of a golden cannon.

If you can tell me the difference between creating images from words and thought, I am all ears.

So, for example, you create a dataset with a bunch of pictures of college football games with them labeled based on the college. From there the program does any number of different math computations (convolution, mainly, but dear god I do not understand it despite taking a class on ML)

A class on image processing would have helped you more. : )

Every global filter you can apply to an image is a form of convolution performed on that image. Blurring, edge detection, sepia tone, whatever. Take two functions/objects and combine them to produce something new.

So when you train a convolutional neural network, it is picking out what convolutions reveal interesting aspects of an image, and permit it to recognize certain things.

You can then invert it and spit out the reverse. 'Apple' points to a number of convolutions, it picks an appropriate set and presents one for your enjoyment. In the case of more advanced models it is clear they do understand it as a 3-dimensional object.

And if it hasn't been trained on that data it just ignores those tokens. I could not get Dall-E to output certain types of hats no matter what I tried.

The datasets that these algorithms use are gigantic and sourced from scraping the internet for art without artists consent and without care for what it grabs. This, of course, has many artists unhappy because it's essentially mass larceny.

Fraud like this is only the tip of the iceberg here. I have also seen a number of 'text generators' that are basically glorified search engines. Some AI art generators could easily be doing what amounts to the same thing is they overfit their input.

[RederReds]

The genie is not going back in the bottle. People need to be aware of what it means and what sorts of discussions need to be had, given the new reality.

To me, AI is a force multiplier. For artists, the end result I see is they will make images and use AI to generate thousands of permutations of each of them. AI art is still very much in its infancy, and 'prompt engineering' can only get you so far with it.

This sort of objection has been made to every advancement in artificial intelligence over the past sixty years. The first artificial neural networks couldn't even perform a XOR operation. Now I can tell one to show me a picture of a chitinous apple being fired out of a golden cannon.

Yup, there's no stopping the advancement of AI, that much I agree with, but I am a firm believer that it can be done ethically. The current way of training these models is sketchy at best and outright illegal at worst. However, part of the reason I believe AI art has been taken so poorly is because it threatens careers for creatives. While I think adapting to AI generators is good and using it as a tool even better, the fact of the matter is many companies and commissioners will see AI generated art as "good enough" and protect their wallet. AI generated art wouldn't be as scary if it didn't literally threaten people's ability to live through their career. Automation is inevitable and we need to adapt but that's a whole other conversation about capitalism, socialism, the virtues of both, etc.

If you can tell me the difference between creating images from words and thought, I am all ears.

Not sure I follow here, apologies. If you want to expand more I'm more than willing to try and elaborate on the point.

A class on image processing would have helped you more. : )

Yeah I mean... It was a Mechanical Engineering class on Machine Learning so take that as you will :P I will not call myself an expert, just someone vaguely aware of the parts under the hood so I appreciate your more though explanation.

And if it hasn't been trained on that data it just ignores those tokens. I could not get Dall-E to output certain types of hats no matter what I tried.

Fraud like this is only the tip of the iceberg here. I have also seen a number of 'text generators' that are basically glorified search engines. Some AI art generators could easily be doing what amounts to the same thing is they overfit their input.

That is the downfall of AI art which is why, for example, we see a lot of the more popular generators stuck in a certain style or some can't draw hands or matching eyes. If they aren't trained on something then it just falls apart. It's the one light in the tunnel for me, personally, however I do believe that at the end of the day we'll see people take it as good enough and/or (hopefully) hire someone to make minor fixes. It's why I still feel secure offering commissions - these generators can't take a "just make the hat red" from a commissioner and tweak a piece the way that's desired. But, I've also seen friends lose their jobs because "AI will do your work :)" which is short sighted and shitty so I have a bit of a personal investment. I don't think these generators will ever disappear, but I would like to see something happen to respect the time and work of artists more.

[Oniya]

That is the downfall of AI art which is why, for example, we see a lot of the more popular generators stuck in a certain style or some can't draw hands or matching eyes.

If you want a laugh (or inspiration for alien life forms), tell one of them to produce a 'hand drawing tutorial'.  I've *seen* some things.

[Vekseid]

Yup, there's no stopping the advancement of AI, that much I agree with, but I am a firm believer that it can be done ethically. The current way of training these models is sketchy at best and outright illegal at worst. However, part of the reason I believe AI art has been taken so poorly is because it threatens careers for creatives. While I think adapting to AI generators is good and using it as a tool even better, the fact of the matter is many companies and commissioners will see AI generated art as "good enough" and protect their wallet. AI generated art wouldn't be as scary if it didn't literally threaten people's ability to live through their career. Automation is inevitable and we need to adapt but that's a whole other conversation about capitalism, socialism, the virtues of both, etc.

And yet you don't see programmers complaining much, which this stuff is already much better at.

It's mostly "I can do so much more now!"

Can automate so much tedium.

That said, there will need to be a conversation about how our economy functions, eventually.

Not sure I follow here, apologies. If you want to expand more I'm more than willing to try and elaborate on the point.

Neural network models are all built off of our own brain. Our brain is still vastly more complex than any current network, but the way data passes through layers of a NN is not so different from thought as we experience it.

Meanwhile

That is the downfall of AI art which is why, for example, we see a lot of the more popular generators stuck in a certain style or some can't draw hands or matching eyes. If they aren't trained on something then it just falls apart. It's the one light in the tunnel for me, personally, however I do believe that at the end of the day we'll see people take it as good enough and/or (hopefully) hire someone to make minor fixes. It's why I still feel secure offering commissions - these generators can't take a "just make the hat red" from a commissioner and tweak a piece the way that's desired. But, I've also seen friends lose their jobs because "AI will do your work :)" which is short sighted and shitty so I have a bit of a personal investment. I don't think these generators will ever disappear, but I would like to see something happen to respect the time and work of artists more.

Well it still breaks down the scene into its convolved components, in theory it should be able to maintain state awareness on an image in the same way ChatGPT does conversations.

Another thing I've noticed with the Stable Diffusion crowd, they have an unhealthy obsession with Emma Watson and I suspect it is a matter of time before legal precedent gets set for that as well.

[RederReds]

If you want a laugh (or inspiration for alien life forms), tell one of them to produce a 'hand drawing tutorial'.  I've *seen* some things.

Oh man yeah that sounds like it would produce some fun results. I wonder if it would replicate the "draw the rest of the owl" meme if you asked it for a tutorial on how to draw an owl.

Neural network models are all built off of our own brain. Our brain is still vastly more complex than any current network, but the way data passes through layers of a NN is not so different from thought as we experience it.

Ah so, as a tangent that might not address your question (which, apologies if that's the case) - I think the distinction in NN versus Artist is the process. While they may process information similarly the process in which art is created and the intent in which it is created are vastly different. Many artistic types tend to be almost spiritual about their work to some degree and I would have to say I'm one of them though minorly so. I enjoy the process and the concept of cutting any of my years of work out feels like I would be sullying what art means to me. The tedium in creating a piece is part of what I enjoy when it all pays off and perhaps that's the masochist in me but I digress.

Really it comes down to the abstract concept of 'soul' in work which there are endless debates on. From a material standpoint there are no differences between art created by a machine and a person, but from a standpoint of soul and intent there are differences. I think that's another part of the vitriol from artists - their art is personal to them. It's a part of who they are. Taking that without their permission is taking something that's a part of their person. That's part of the difference between the reaction programmers have versus artists, I think, but I could be very wrong. I've done both and I absolutely understand the appeal of taking tedium out of programming, but I've also only programmed lightly for a few years so I am nowhere near an expert on the subject.

Another thing I've noticed with the Stable Diffusion crowd, they have an unhealthy obsession with Emma Watson and I suspect it is a matter of time before legal precedent gets set for that as well.

Another thing I've seen is people suggesting people train a model on Disney's work as a way to force some sort of legal precedent. Any model doing that would get slapped down real fast by the big mouse.

[Kuroneko]

Please excuse the flip meme.

I snipped this bit of RederRed's post because this is also part of what I'm seeing from artists, and it's also something that resonated with me.

Really it comes down to the abstract concept of 'soul' in work which there are endless debates on. From a material standpoint there are no differences between art created by a machine and a person, but from a standpoint of soul and intent there are differences. I think that's another part of the vitriol from artists - their art is personal to them. It's a part of who they are. Taking that without their permission is taking something that's a part of their person. That's part of the difference between the reaction programmers have versus artists, I think, but I could be very wrong.

I can't speak for programmers either, beyond saying that I'm not sure it's an apples to apples comparison. Programmers and artists likely have different attitudes towards these issues, but maybe I'm assuming.  I personally can't see why would I want an AI program to do my art process for me. It's the process of creating it - and everything that goes into it, which sometimes include longs hours - which makes it enjoyable in the first place. It's the difference between buying a meal and cooking it yourself, I think. Can I buy a better meal than one I might make myself? Sure. But buying a meal will never replace making one of my mom's favorite recipes, or one that we used to make together as I grew up, where she taught me how to cook, season, and prepare a meal. There's more personal investment in the latter. The sense of accomplishment and satisfaction that comes from that isn't the same as buying dinner. Creating a work of art is the same, at least for me. I don't get that same experience from stringing together a set of prompts to feed into Midjourney. Maybe it's not the best analogy, lol. Of course, I'm only speaking for myself. I'm sure that there are plenty of artists who would love to cut down on some of the repetitive tasks that their work involves. I know lots of other costume designers that use their previous drawings of human figures for their renderings, rather than draw out new ones to speed up their process, for instance.

Other professors I know are now seeing papers written with chatGPT turned in by students, which means we'll soon have to come up with language in our syllabi and grading rubrics to address the use of AI text generators by students, as well as ways to determine if something was AI generated. I anticipate we'll be discussing it quite a bit at my university.

[firepyre]

I don't much think this will be a popular viewpoint amongst artists, but from a layperson's perspective, having access to AI tools that we can use to spit out art-like images is fantastic. No need to deal with copyright, and even someone with minimal skills can get a pretty decent looking result. For most uses where I just want to convey a concept or idea, I don't really care if it has soul or not.

On the other hand though, or something you hang on your wall, or in a gallery, something intended to be the center of conversation, then intent does matter. There is a place for art created with intent, and I think AI largely lacks that. That said, there has been art competitions won by AI compositions, so maybe I place too much importance on that. As the beholder, if I can't read into the artist's intent, then does it really matter if the piece was created by a machine or a human?

While it sucks that some artists might lose their jobs, that's really no different to any other industry -. In my opinion, neural networks have the potential to shake things up on a level similar to that of the industrial revolution. We've probably only seen the very tip of the iceberg yet. And like any big change, there's going to be a whole slew of new ethical and social issues to deal with, and a bunch of growing pains while society sorts out some kind of new normal. Exciting and scary in equal measure.

[Oniya]

The thing is, with the way the programs 'learn', there's no way to opt-out.  You upload prom pictures, and they could get scraped and conglomerated into one of those 'flight attendant' pictures that Reder mentioned.  Emma Watson might have access to good lawyers, but what about Judy Smith?

[GloomCookie]

My opinion is simply this. AI artwork does scrape from images, but then... everyone does. How do we not learn than from observing the classics and giving our own interpretation to it? For example, this:



Now I ask, is this one any different if it was made with AI or by a human hand?



The point I'm making is that all art is built on something else. What's the cutoff before you can claim that it's no longer an original piece and is just an amalgamation of other works? Is there a percentage of 'original' art that needs to exist? If so, does that 'original' have to be so unique that it has never existed before or since in our universe?

Also, there's a lot of artists mad that they can't charge as much for commissions because someone can put together an AI portrait of their D&D character. That's legit, but then not everyone wants to spend $50 for a character they might get to play three times before the game dissolves. And if it does go longer, so what? They have a representation of their character. If they want to bad enough, they can then take the AI image to an actual artist and say "Here's my character's base image, can you please draw them doing this epic thing?"

I see AI art as a tool, nothing more. It's useful for me to create something to tell the story, such as a beautiful city or a dark castle, something I can throw together in 5 minutes with some level of consistency (which is a huge deal for me) without having to do the same thing on Deviant Art or Google or Art Station for over an hour finding just the right picture that matches the art style I have going. I can instead type a few parameters into the AI program, get a few images that are fairly uniform in appearance, and gets the point across better.

And finally, I use AI to create a custom avatar for myself, and I don't have a fortune to spend on having every possible emotion or clothing combination just to put out a little blurb about how excited I am or that I'm munching breakfast with the family.

Or holding a kitty.

[Kuroneko]

There is no getting around the copyright issue. The issue is there whether we want to ignore it or not. I have no issue with the technology at all, only the way images have been taken without permission or compensation. As mentioned, the tech is here. We now have to find a way to use it ethically. Has an AI piece won an art competition? Yes, at a state fair, which is not exactly the pinnacle of art competitions. However, I have no doubt we'll see AI submitted to competitions with more reputation, leading to more discussion surrounding its use.

Now I ask, is this one any different if it was made with AI or by a human hand?

I'm not sure that the Mona Lisa and a derivative image is the best example, for a coupe of reasons. One, the Mona Lisa itself is in the public domain (though photos of it may be under copyright). No artist will have their income, potential income, or reputation affected by AI or an artist inserting a new face into the image, or making otherwise derivative works of it. Two, are these images any different? Astonishingly so, in my opinion. While the second image is beautiful, it has none of the characteristics that make Da Vinci's work a masterpiece. There is no chiaroscuro, no sfumato, the very things that made him the master he is, the things that make the Mona Lisa the masterpiece it is. To me, aside from the obvious copying of the composition with a new face and background, the second image looks more like Edward Burne Jones' style. But, his work is also in the public domain. So, no harm, no foul. Whether the second image was made by a human or AI, it's a blatant copy of Da Vinci's work, so I personally can't find the same appreciation for it as I do for the original. If a human made it, I could see the value in it as a learning exercise, but not as a final piece for their body of work.

The point I'm making is that all art is built on something else. What's the cutoff before you can claim that it's no longer an original piece and is just an amalgamation of other works? Is there a percentage of 'original' art that needs to exist? If so, does that 'original' have to be so unique that it has never existed before or since in our universe?

I don't think amalgamation isn't quite the right word. Do artists hone their skills by learning the techniques of those that have gone before? Do we do copy work as exercises? Absolutely. I'd never sell them as my own work, however, ore represent them as anything more than a study. I recently took a watercolor class from one of my favorite artists, Aria Fawn, where she generously taught her painting techniques. Did any of my work from that class look like hers? Not at all. I don't want to copy her work, I want to improve my own. I don't know a single artist who wants to make their work out of an assembly of other artists' work. Instead, we look to interpret the world around us in a unique, individual way. There's no need to do the Mona Lisa again; it's already been done. We want to make something new.

Artists aren't mad that they can't charge a set fee. They're mad that they're losing jobs, and that the work of artists in general is being devalued. They're mad that their work was used to train a program without their permission. Their mad that filters are being created using their signature style, which was the case for Hollie Mengert. But you know what? Artists deserve to get paid for their time, like any other worker. So, if it takes me an hour to do a finished piece, I'm going to charge the client for that hour. If it only takes me ten minutes to make a sketch, I'll charge accordingly. If I used an ethical AI tool to generate variations of one of my own sketches and it took five minutes, I'd only charge for five minutes.

By all means, use the tool. I understand the appeal. Let's make sure it can be used ethically.

[HannibalBarca]

The first semester I went to college, I worked for a newspaper.  This was 1988-89, and the printing press was a newer version that had a computer connected to it, so that the pages could be type corrected on the computer and fed directly into the press.  Before that, a typesetter had to arrange metal type in rows, by size and font...by hand.  Yes, just like Gutenberg did hundreds of years before.  When the computer was combined with printing presses, it was a huge leap forward in printing technology.

Now that first semester, my major was photography.  My grandfather had worked with me in his home darkroom, and I knew how to mix chemicals, develop film and prints, enlarge them, even do some tricks like burning and dodging to make double exposures and such.  During the semester, our photography teacher took us over to the computer lab, to meet a professor who was working with something cutting edge--digital cameras.  I'd heard about it, but seeing someone simply pull out a data card and stick it into a computer, then call up the images they just took--it was stunning.  Even more printing them out, even if at that point printers couldn't match photo paper pixel quality.

It was the talk the computer lab professor gave us that changed everything for me.  He told us the change was coming, where not only digital photography, but computer imaging and processing, was going to do away with thousands of jobs.  He told us how years before, he had to go to representatives of the typesetters union, who worked at newspaper and magazine companies, and tell them computers were going to do away with their jobs.  They laughed at him.  Five years later, there were no typesetter jobs left, anywhere.  He said that the same thing was going to happen to darkroom techs and photo finishing.  Computers would do all of that.

I was gobsmacked.  I'd been planning to be a photographer, but was currently employed at the newspaper as a darkroom tech.  I wasn't a good enough photographer or journalist yet to have a full-time job with one of those, and I took him seriously.  So I changed my major to desktop publishing, computer reprographics, and computer animation.  That professor wrote the curriculum for those degree tracks for California junior colleges, and I earned two of them from taking his courses.  I worked on Photoshop--back before there was a 1.0, 2.0, or so on.  Every day he'd tell about a new effect on the program he'd learned from last night's experimenting on it.  I learned the basics of video editing, multimedia...so much of it was just getting off the ground.

And I lost my job at the newspaper as digital cameras became the norm.  Not because they replaced me or fired me--because the newspaper went under, when their costs with traditional film and print and other costs were too great compared to the other town paper, that went digital, saved money, and lowered the price of their paper accordingly.  I, therefore, have first-hand experience about technology coming in a wiping out your job.  It's ironic that I was learning the new technology--but I couldn't find a job at first that would hire me to use it, because I lived in a small town, and wasn't willing to move to a bigger city, where I could get an in-demand job with an advertising company, television network, or magazine.

I'm a visual artist as well as a musician and writer.  I've done pencil/pen/graphite illustrations for decades, and am a competent portrait artist.  I've also begun using MidJourney, the AI art generator, too.  I can see the same issues here that I saw when I was a young man, and the same issue that has cropped up for millennia, when technology leaps forward.  Blacksmiths lost jobs when automobiles were invented, telegraph operators lost jobs when telephones were invented, kamishibai men lost jobs in Japan when televisions were invented--there are countless stories of history running roughshod over entire career fields when time moved on.  Humanity is relentless to itself when we develop new stuff.

Where it becomes brutal is when money becomes involved.  I don't use pictures I create for income--only for pleasure.  I already think the works of dead artists should be in the public domain, so if I make an image in the style of Frank Frazetta it doesn't cause me any anguish.  For those who are alive--I wouldn't sell anything in their style, if I do use it.  But--I wonder if simply paying money to an AI generator's creator is feeding the beast by rewarding them for copying art I don't have the skill to replicate in my own hand, especially with me being older now, and losing some of my manual dexterity.  By simply being a functioning member of society, I am already assisting in the continuation of so much of the end-stage capitalism I already despise, but life has to be lived as you can, when you're not a plutocrat or even member of the comfortably middle class.

I don't want artists to lose incomes on work they deserve to have, but at the same time, how much of it is inevitable, like what happened to me when I was younger?  And what happens when software and hardware can do the work of what 95% of us in the general population can accomplish?  Machines already run assembly lines, grow food, manage electrical grids, organize supply lines, direct traffic on roads, skyways, and railroads...will it come to the point where the only jobs necessary by humans are those of repairing the computers and machines that do all the other work?

I do know, however, that much of this is driven by the profit incentive, which boils down to someone wanting more than what others have.  We need to get a grip on letting the sociopaths and narcissists into positions of power and control, and spend less time focusing on how the little people spend their free time making images they couldn't afford paying a professional artist to draw, anyway.

[Timeless]

~snip~

I'm gonna say this, but this post really disheartened me. Kuroneko basically nailed what I wanted to say, but I do want to link a video which basically explains how the artists, including me, feel about this.

Why Artists Are Fed Up with AI Art

https://youtu.be/5Viy3Cu3DLk

[Dice]

I hate that photos and images (And as a photographer I am sure my work is in this mess somewhere) are bluntly stolen wholesale to make this shit happen.

I get that there is no going back, but how we got here was bullshit.

[Rinzler]

I would never of course suggest that hackers and trolls might find some way to flood these datasets with nazi iconography. That would be despicable. Even if the resultant shitstorm would be very amusing to behold.

In any case, it does raise an important point, I think: that these datasets may very well be subject to the whims of the people who curate them. Hardly a reassuring thought, given the present impetus of AI art.

[Chulanowa]

My opinion is simply this. AI artwork does scrape from images, but then... everyone does. How do we not learn than from observing the classics and giving our own interpretation to it? For example, this:



Now I ask, is this one any different if it was made with AI or by a human hand?

Couple key differences.

1. Leonardo da Vinci has been dead for five hundred and three years.

2. There is literally no one connected to modern media on this planet who will not immediately recognize the latter as a spoof of the original by name and creator.

"AI Art" is stealing from extant artists trying to make a living, using their own art and style against them, forcing them to compete against a theif with a print button.

[GloomCookie]

The concept of copyrights are disturbing in their own right. Mickey Mouse was created in 1928, yet the Disney corporation still owns the rights to him exclusively. The original creators are dead, have been for decades. Yet a corporation owns the concepts. The way Copyright works is that Star Wars will remain under copyright until 2072 at least.

Yet if I patent an idea, which is similar to copyrighting but for concepts that I can build and manufacture, If I patented an idea the moment I hit post, it will be available to anyone who wants to use it on December 27, 2042.

Why is one type of idea more protected than another? If da Vinci's work was still under copyright, how would you feel?

OR! Let's consider something else.



This image is 100% legal.



This image is not.

What's the difference? French law says that architecture is protected under copyright. The original design for the Eiffel tower is no longer under copyright, yet the lights were installed later and are still protected by copyright.

Copyright law is a giant mess and full of complications, contradictions, and idiocy.

I get it, protect your work for a period of time, but why is it that copyright law on works and images are so insanely long? Would it be ok if all the images an AI used datasets from were over 20 years old? Or do they all need to be public use and not covered by a copyright that is arbitrarily long? What's a reasonable use expectation if the AI is adapting and changing the work instead of just copy/pasting like thousands of people currently do? You post ANY image on the internet, it WILL be posted somewhere without your copyright info and will be distributed far and wide and you won't see anything for it unless you got paid upfront.

Try doing a reverse image search on Google for any random image you find. I doubt you'll find anything at all. Thousands of images are lifted from their source, so trying to credit every single artist is impossible. Trying to track down artists to do 'fair use' and 'attribute credit' will open up the developers to insane legal headaches, especially since AI might pull from random parts of one image, parts of another, and there's no telling where it will come from.

[Oniya]

*drags the goalposts back into the stadium*

The issue of corporate copyrights is quite a bit removed from the issue of artists who are alive, or people who post pictures of themselves/family members on Instagram, or medical records that shouldn't have been able to be scraped in the first place.

[Kuroneko]

Mickey Mouse as an image will actually go into the public domain on Jan. 1, 2024. However, the Walt Disney company will still own the trademark and copyright of the character's name, along with thousands of other symbols that are connected to or associated with the character, and those protections will last as long as the company uses Mickey Mouse as their company logo. So, should anyone decide to use the image of Mickey that goes into the public domain in a year, it can't be used in any way that implies any kind of association with Disney, or they'll face trademark violation. But as Oniya said, corporate copyrights are quite a bit removed from the issue at hand (and are far more complicated).

Copyrights exist for as long as they do because they seek to protect the author's work, not just for their lifespan, but to ensure that their heirs will receive the benefits of their work. Ownership of a copyright is a lot like owning shares in a company. Copyright terms provide some measure of 'inheritance.' This article explains it well - http://copyright.nova.edu/copyright-duration/

If Da Vinci's work was still under copyright I'd hope that his descendants were receiving at least part of the enormous amount of money that's made from the mass production of his work on everything from posters to coffee cups. I'd hope the same for Van Gogh's.

It's remarkably easy to reverse Google search an image. I did it for the Eiffel Tower images posted here and found their sources immediately (Wikimedia commons, and Upsplash respectively). I think there's some confusion about what makes an image 'legal' or not. The legality and fair use of the Wikimedia image has nothing to do with the fact that the copyright on the tower's architecture is in the public domain. It can be used freely because the image itself is is in the public domain, because the user who uploaded the image shared it under the Creative Commons Attribution-Share Alike 3.0 Unported license. Likewise, the second image isn't technically illegal because the photographer isn't selling it for commercial use. You can download it for free on Upsplash. While Paris doesn't want commercial photographers taking pics of the tower at night, they don't go after tourists. https://www.travelandleisure.com/photography/illegal-to-take-eiffel-tower-photos-at-night

It's also incredibly easy for artists to search for their own work on haveibeentrained.com to see if their work was used by Stable Diffusion.

[rayguncourtesan]

This image is 100% legal.



This image is not.

What's the difference? French law says that architecture is protected under copyright. The original design for the Eiffel tower is no longer under copyright, yet the lights were installed later and are still protected by copyright.

Copyright law is a giant mess and full of complications, contradictions, and idiocy.

That...is not how copyright works. The image is perfectly legal (subject to the license granted by the photographer, who has a copyright over the image). The architecture copyright protects the architecture; you cannot arrange those particular lights in those exact positions. The image is its own thing; created by the photographer who took it and potentially disposed of to a purchaser for valuer.

To properly frame your example we would have to imagine that I have, compliant with copyright law, constructed my very own Not!Eiffle Tower. It is exactly the same as yours in every respect. This is legal because the Eiffle Tower is not under copyright.

You have put lights on your Eiffle Tower (pictured). These are protected under copyright in French law; I cannot put the same lights on my otherwise identical tower. You created something new, a work of architectural art and I cannot steal it from you.

That is copyright. I do not have the right to copy it. I can put my own lights on my tower and I can even be directly inspired by yours, but they must be sufficiently different to not contravene your copyright.

I can of course enter into negotiations with you over the copyright you have over your lights. I could offer to pay you to duplicate them and you have the right to accept or reject my offer, or to offer terms under which you would be prepared to allow them to be copied.

That is the purpose of copyright; at its most basic, it is to incentivize the production of art on the basis that if I spend all my time protecting my stuff from you, I do not have the time to produce new stuff. It is an extension from physical property because an idea is not diminished by its reproduction; if you have an apple and I take your apple, you no longer have an apple. If you have an idea and I take your idea...you still have it. But if you cannot profit from your idea, why bother having them?

That's where AI art and the basis of copyright law come into conflict. Why should artists continue to produce art if it will be immediately stolen by AI's? AI's cannot produce from nothing; their work is composite and iterative but it is not inspired. A world in which AI is the only producer of art will mean radically different art to a world in which people produce art and if something is not done to protect the work of artists, that is what will happen. Artists will stop producing, the datasets will become recursive and since AI operates on trillions of iterative generations, we will end up in a place of mono-art.

Capitaism after all produces not what is best, but what is most profitable to corporations.

Copyright law offers the best defense against a rather bleak cultural future and I wish more artists had a firmer grasp of what it was and how to use it.

[Vekseid]

That...is not how copyright works.

In Italy, France, and other countries that have opted out of freedom of panorama, that is exactly how it works. Even in the US, you can get in trouble if you make a statue or other non-inhabitable artistic structure the focal point of your photograph.

In the US, however, second picture would be completely legal as it qualifies as architecture since it has facilities for humans.

[RedRose]

I learned recently that some think taking pics of a human is forbidden! Around here, schools post pics without asking.

[rayguncourtesan]

There is a world of difference between commercial and non-,commercial use of an image and pretending that holding a copyright and enforcing it are the same thing is disingenuous at best.

And yes, law is such that a lot of things are legal in some countries and illegal in others but if we're discussing enforcement of copyright (mere holding is meaningless) then the issue is where the enforcement shall take place.

[GloomCookie]

The key is if this will be enforced. Even if an artist goes out of their way to sue Stable Diffusion or whoever is making the AI program, they could very well use the argument that because the images are being manipulated and thus are no longer the copyright images that were originally protected. And if they happen to come out identical, they can try to point out the AI is using user input to craft an image, and therefor they are not liable.

Basically, they can (and probably have) gone at this from a legal angle to make sure they don't pour a ton of work into this AI program only to have it blow up in their face because someone sued them about their image database.

[Kuroneko]

If an image comes out identical, that would be proof an artist's image was stolen, not that the program randomly produced an identical image with user input.

The effectiveness of prosecution remains to be seen. I wouldn't be surprised if these image scrapping companies never even thought about copyright, in this day and age when everyone thinks that if it's on the web, it's free.  Personally, I hope Disney goes after them. The company is notoriously litigious.

The Concept Art Association is gathering funds to protect artists from AI technologies on GoFundMe as we speak. MeFu - Mestieri del Fumetto - is as well. Both are working to create new legislation and regulation of the practices these companies use to collect images in the U.S. and the EU.

[TheGlyphstone]

Over at SMBC, Weinersmith points out that when automation takes over production, the handmade stuff becomes 'artisinal' and gets sold to rich idiots at 20× the price.

[Keelan]

Over at SMBC, Weinersmith points out that when automation takes over production, the handmade stuff becomes 'artisinal' and gets sold to rich idiots at 20× the price.

Sure... when people are assuming that you ACTUALLY made the work yourself. Which is going to be a significant problem because... well, I've paid for about two dozen art pieces in the last five years, and ALL of them have been digital artwork which is the medium AI works in.

I mention this because I stumbled upon this today: https://nichegamer.com/art-subreddit-bans-artist-style-ai/

The work in question was this one: https://www.reddit.com/r/drawing/comments/ztbtw9/a_muse_in_the_warzone/

Artist is Ben Moran, and while I won't link his artstation or deviantart here, his art style is like this and he has years of work in the style.

[TheGlyphstone]

Oh, it won't be as easy. But if the idea that all AI-made art will homogenize into a single style is true, then it's possible for an artist - granted, not that specific dude - to have a style very distinct from the 'vanilla' flavor of AI. And unless they're Andy Warhol 2.0, it's unlikely their output being scraped will shift the vanilla-AI art algorithm enough to be meaningful.

This still sucks for artists in the now, 120%. I'm just throwing a marker of skepticism on the idea that this will kill the idea of original art for all time.

Plus - could any human mind conceive of this wondrous horror?
https://pbs.twimg.com/media/FlkKaO3aUAAtxx_.jpg (Linked for kidlets)

[Kuroneko]

Plus - could any human mind conceive of this wondrous horror?
https://pbs.twimg.com/media/FlkKaO3aUAAtxx_.jpg (Linked for kidlets)

Human minds have already conceived of far worse, while being far more original. Horror versions of children's characters are nothing new, as a simple Google search for 'scary Disney princesses' shows.   

[TheGlyphstone]

Not my point, by a million miles. I've seen plenty of scary children's characters and evil Disney princesses...there aren't even any princesses in that photo. What those pictures have that I've never seen in any human-made art is the simple inhuman grotesqueness. AI does not understand human anatomy or what a person is supposed to look like, and that simple incomprehensibility of something any human brain instinctively grasps (even if they then choose to subvert it) resulted in what I will say is a unique creation.

[Kuroneko]

Not my point either. The google image search was simply an example, not a  reflection of what you posted. I work in film and theatre, specifically special effects makeup, and I have seen grotesque images like that created by humans a million times. It's not unique to AI by a long shot.

[Kuroneko]

Not my point either. The google image search was simply an example, not a reflection of what you posted. Obviously there aren't any Disney princesses in that photo. I work in film and theatre, specifically special effects makeup, and I have seen grotesque images like that created by humans a million times. It's not unique to AI by a long shot, and human minds are quite able to grasp it.

[Kuroneko]

Sorry, my cat ran over my keyboard and accidentally quoted me mid posting. *sigh*

[Oniya]

Oh, it won't be as easy. But if the idea that all AI-made art will homogenize into a single style is true, then it's possible for an artist - granted, not that specific dude - to have a style very distinct from the 'vanilla' flavor of AI. And unless they're Andy Warhol 2.0, it's unlikely their output being scraped will shift the vanilla-AI art algorithm enough to be meaningful.

This still sucks for artists in the now, 120%. I'm just throwing a marker of skepticism on the idea that this will kill the idea of original art for all time.

Plus - could any human mind conceive of this wondrous horror?
https://pbs.twimg.com/media/FlkKaO3aUAAtxx_.jpg (Linked for kidlets)

*glances at upper right image*  I'd like to introduce you to this unknown indie game called 'Five Nights at Freddy's'.

[Kuroneko]

*glances at upper right image*  I'd like to introduce you to this unknown indie game called 'Five Nights at Freddy's'.

Or Jorge Dos Diablos...
...or Guillermo De Toro ... or Heironymous Bosch ... or Carpenter's The Thing ... or American McGee's Alice ... or Christien Van Minnen ... or Gary Tunnicliffe ... or Georgia Straight ... or H.R. Giger ... or ...

[Rinzler]

Getty Images is suing Stability AI - the folks behind Stable Diffusion - alledging unlicensed use of its image libraries.

https://edition.cnn.com/2023/01/17/tech/getty-images-stability-ai-lawsuit/index.html

Colour me very curious to see how this goes. If Getty wins, the the AI art tool creators will probably have to be a lot more circumspect in their image scraping practices. Which'll no doubt reassure a lot of artists out there.

[Dice]

This is likely to end up being a major issue for groups that have AI art programs. If your TOS claim they own the copyright to the "Art" they make they are in for hell over this.

https://petapixel.com/2023/02/02/ai-image-generators-can-exactly-replicate-copyrighted-photos/

TLDR:
You can get AI programs to spit out a near 1 to 1 of the images they were trained on. So close they would be in legal shit if someone tried to use them.

[Kailandra]

One of the key issues with machine learning that I don't see being discussed enough are the ulterior purposes.

Let's just start with the relatively benign. AI art wasn't a bid thing until Elon purchased Twitter. He was quickly caught up in demands from artists to remove unauthorized used of their artwork, or get royalty payments for their art being displayed. All of a sudden, we've got AI art, which anyone can use and has no copyright attached. No more hassles over royalties and DMCA.

The scary part: advances in machine learning have come in support of government survielence. China was having problems with their facial recognition software identifying masked protestors. Next thing you know, Ten cent has a free app to 'see what you look like as an anime character. Fun, but also providing fodder to refine the facial recognition system. Similarly, the FBI was having difficulty with facial recognition in cases where there was heavy, contrasting make-up being worn. Oh, hey! A fun new camera filter to see what you look like as a Juggalo...

On a tangent, AI chat bot have undergone a maked improvement and are becoming more popular. It's entirely possible that there could be new apps on E being written by these programs in the near future.

[Oniya]

The scary part: advances in machine learning have come in support of government survielence. China was having problems with their facial recognition software identifying masked protestors. Next thing you know, Ten cent has a free app to 'see what you look like as an anime character. Fun, but also providing fodder to refine the facial recognition system. Similarly, the FBI was having difficulty with facial recognition in cases where there was heavy, contrasting make-up being worn. Oh, hey! A fun new camera filter to see what you look like as a Juggalo...

Of course, machine learning is quite dependent on the 'teachers'.  CAPTCHA programs are used to train self-driving cars, for example, which is why so many have urban/roadway related themes.  Then you end up with something like this:



Guess what it thought was the other 'parking meter'?

[Humble Scribe]

Like XKCD's comment about Self-Driving Cars being more like 'Crowdsourced Steering'.

[Oniya]

Mind you, if someone wants to put money in my mailbox, I'm all for it.  ;D  (I'd have to get one to put at the curb, though.  Unless a bunch of y'all can select a door-flap as a parking meter?) 

[Keelan]

Thought y'all might find this interesting: https://www.theverge.com/2023/2/22/23611278/midjourney-ai-copyright-office-kristina-kashtanova

The original letter/primary source is here: https://www.copyright.gov/docs/zarya-of-the-dawn.pdf

[Oniya]

I have to admit I found this bit of the brief amusing:

The Urantia court held that “some element of human creativity must have occurred in order for the Book to be copyrightable” because “it is
not creations of divine beings that the copyright laws were intended to protect.”

[Keelan]

https://paizo.com/community/blog/v5748dyo6si91

Over the last few months, the world has seen a huge upsurge in interest, use, and quality of algorithm-generated imagery and text. Since we launched the company in 2002, Paizo has made its reputation with the assistance of countless traditional artists and writers, who are just as integral to the success of our games as our in-house editors, art directors, designers, and developers. The ethical and legal issues surrounding “AI art” and writing prompt programs—and the serious threat they pose to the livelihoods of partners who have helped us get to where we are today as a company—demand that we take a firm position against the use of this technology in Paizo products.

In the coming days, Paizo will add new language to its creative contracts that stipulate that all work submitted to us for publication be created by a human. We will further add guidance to our Pathfinder and Starfinder Infinite program FAQs clarifying that AI-generated content is not permitted on either community content marketplace.

Our customers expect a human touch to our releases, and so long as the ethical and legal circumstances surrounding these programs remains murky and undefined, we are unwilling to associate our brands with the technology in any way.

Stated plainly—when you buy a Paizo product, you can be sure that it is the work of human professionals who have spent years honing their craft to produce the best work we can. Paizo will not use AI-generated “creative” work of any kind for the foreseeable future.

We thank the human artists and writers who have been so integral to our success in the past, and we look forward to working with them for many years to come.

[Kuroneko]

https://paizo.com/community/blog/v5748dyo6si91

<3

[Sludgewave]

The big problems I see are:
1) It'll put artists out of work, unless they can make it in the "get rich people to pay for your human art as a luxury". (Which requires there to be rich people with lots of luxury income in the first place).
2) People who have control of the AI will become wealthy off its use, when all they do is own it.
Really, just shows the real core problem is capitalism, which itself is something to fight against for a myriad of other reasons.



As far as the terrors of AI go, I'm more worried about the implications of its use in disinformation/propaganda, or the idea that once we create Skynet it'll be all transphobic or something.

[Lilias]

This is about chatbots, rather than art, but I think it nails the problem with AI.

https://www.christiancentury.org/article/features/problem-artificial-intelligence-us

[LemonMelon]

In general I'm not against AI-generated artwork—I'm even moderately invested into it myself—but in my opinion it kills most incentives for humans to pursue a career in art. I'm certain there are many would-be artists who enjoy drawing/painting or even doing character 3d modeling, but who at the same time couldn't justify investing many hours daily if they didn't have at least the slim hope that at some point in the future they could earn a living with their skills, doing what they like. Much of this is now replaced even at a professional level with rather boring "prompt engineering". If one's goal of making art is just obtaining a pretty picture, certainly AI-generation allows to reach "close-enough" results in a fraction of the time and costs.

A second related problem I have with it is that there's not really much to be learned in terms of fundamental skills with prompt engineering. Furthermore, the current interfacing methods for generating AI artwork will certainly change in the near future, potentially making soon useless again the time invested into learning what effect words (or their combinations) on the final output.

[Vekseid]

Prompt engineering is a kind of funny concept. It exists because the AI still doesn't possess a genuine model of what it is creating. Words are loosely associated with objects presented in its training data, but it doesn't actually have an inherent conception of that object, and there is a lot of noise in the training.

And if what you want isn't in the training data it becomes rather useless.

Stuff is going to start getting crazy (and possibly scary) once these things are used to build models instead.

[Dice]

Here's a short post I wrote a bit ago about this from the creators side and my views on it, call me old fashioned, but I think this true.

I can't draw, at all. But I still partake in a form of art. I sat in the darkness, in a forest for hours to take three photos. (Many photos, stacked. End product was 3 images) If you can do that in a few seconds with a touch of your keyboard you don't get to understand what it is to build a craft.

There is no soul is something that is made by throwing darts at a board and keeping the one you like. There is nothing like working out ahead of time what you plan to show, framing, taking steps towards a goal and then learning from your mistakes. Adjusting your skills in the future.

AI art's true flaw is that it asks so little from the "creator" and thus leaves very little in the ways of development, self reflection or healing that other forms do. Asking someone to sit alone with their thoughts for hours innately requires time for self reflection, understanding and care. Prompted by one's art is more meaningful then prompt and pray.

[GloomCookie]

Framed in that way, I can understand some of the reluctance by artists to accept AI art. It's similar to my skills as a designer/engineer. We have a lot of tools at our disposal, and we're developing new tools to try and make our jobs faster, but they'll never replace the actual person, at least not for a while.

I get it though. There are a lot of designers out there who are so overworked that they'll take any shortcut they can, including having some tools that just drop their work in and call it good, without having to do much more than make a few connections. And for some situations, that's perfectly acceptable, since they're routine and even boring because they do the same thing time and time again. But, the program we're developing doesn't know what to do after that, and that's where the designer must step in.

What I think will end up happening is that a lot of people will look at AI art as "Good enough" for their average purposes and then when they need something that AI can't do, they'll start turning to the real thing. If, say, the local university wants to add some people to make their brochure look better (That happened once at the University of Wisconsin) then AI might be good enough. But if you want something super specific, then you'll need to get someone who is capable of creating actual art.

That's kinda how I see things. Where I think AI will flourish is as a personal assistant, able to start custom tailoring recommendations for specific applications. For example, a lawyer might use it to look up similar case law, or someone like me might use it to research codes and similar projects. I doubt we'll ever get rid of AI given how much money's been poured into it at this point, but we can certainly curtail some of its more negative aspects.

[Al Terego]

That's kinda how I see things. Where I think AI will flourish is as a personal assistant, able to start custom tailoring recommendations for specific applications. For example, a lawyer might use it to look up similar case law.

Why not allow LLMs custom-trained on legal data to substitute lawyers?  After all, if ChatGPT4 can pass the bar exam in the top 10%, why must one pay a (statistically inferior) lawyer $400/hour, when they can get (statistically better) results for $0.12/1K tokens?

Oh right, because the legislators will ban that use to protect their caste.

[Keelan]

Why not allow LLMs custom-trained on legal data to substitute lawyers?  After all, if ChatGPT4 can pass the bar exam in the top 10%, why must one pay a (statistically inferior) lawyer $400/hour, when they can get (statistically better) results for $0.12/1K tokens?

Oh right, because the legislators will ban that use to protect their caste.

ChatGPT can't represent me in court as far as I know. It also can't help out at all during trials.

ChatGPT also doesn't necessarily know when and what to file or how the filings have to be formatted and bound (yes, that does matter for some courts), nor necessarily how many copies you're required to file.

ChatGPT can write up a contract, but if you don't know how to read contracts you can end up fucking yourself over and words in law often have very specific meanings.

ChatGPT can write up a contract or any legal document, but are you 100% sure that a document it created is in fact accurate? Do you know enough about the law to be able to double-check the ai?

ChatGPT would potentially require intimate knowledge of evidence potentially which would be a MASSIVE legal issue since in order to feed it to ChatGPT I'm pretty sure you run into evidentiary issues and can potentially ruin your case.

My point being that as much as lawyers tend to rub people the wrong way - and I've worked for lawyers before - they do actually provide a useful skillset in dealing with legal systems. This isn't to say that ChatGPT won't have utility - it absolutely will - but I doubt that we'll be able to completely price them out.

[Al Terego]

I specifically wrote "LLMs custom-trained on legal data".

And the fact that it "can't represent me in court [and] can't help out at all during trials" was the whole point of my post.

[Oniya]

And the fact that it "can't represent me in court [and] can't help out at all during trials" was the whole point of my post.

It's really hard to get that from your prior post, which stated:

Why not allow LLMs custom-trained on legal data to substitute lawyers?  After all, if ChatGPT4 can pass the bar exam in the top 10%, why must one pay a (statistically inferior) lawyer $400/hour, when they can get (statistically better) results for $0.12/1K tokens?

Oh right, because the legislators will ban that use to protect their caste.

It's not that ChatGPT isn't allowed to represent someone in court, it is that it is functionally unable to effectively do so, due to the other four points that Keelan made.  I would like to see an actual source stating the 'statistics' that you loosely reference here.

[Keelan]

I specifically wrote "LLMs custom-trained on legal data".

And the fact that it "can't represent me in court [and] can't help out at all during trials" was the whole point of my post.

And my point was that if you're not an LLM (Master of Laws) and you're using an LLM (Large-language model) for legal matters then there is potential for a lot of problems that can be caused by one's ignorance of law and legal procedure, or alternatively one's stupidity with regards to how they believe the law *should* work.

Also the stuff relating to confidentiality, evidentiary protocols, disclosure policies, privacy, security, etc.

And while your whole point seemed to center upon the "Legislators [banning] to protect their caste", my point of why it wouldn't work stemmed largely from the fact that ChatGPT is not a legal person and has no legal standing. I cannot secure ChatGPT as counsel not because they want to 'secure their caste', but because my counsel has to be a legal person registered to practice law in the state and in that area of law.

IF ChatGPT were able to be used in the courtroom, you would legally be representing yourself as you do not have counsel representing you. This means that Average Joe with ChatGPT is up against Bigshot Lawyer with ChatGPT in court, or worse yet Bigtime-Election-Year Prosecutor with ChatGPT (because why wouldn't they use that tool too?). You likewise then couldn't appeal the results due to ineffective counsel because you never secured counsel and were representing yourself.

[Al Terego]

Reply to Oniya

Hmmm...  Alright, I'll attempt a better presentation.

Fact: OpenAI claims that ChatGPT4 scored in the top 10% of the Uniform Bar Exam (source: Forbes).

Conjecture: If we assume that a lawyer's score on the bar exam has some correlation with how "good" of a lawyer they are, it follows that ChatGPT4 is "statistically better" than most lawyers.

Now, I do realize that experience plays a large role, and that's why the rest of my comments below will not refer to ChatGpt4 specifically but to a hypothetical LLM extensively trained on legal data.  Given the huge amount of court transcripts, case law, and other documents that is available, this is quite doable.

Assertion: My personal opinion is that there isn't anything that a defence lawyer can do that a specialized "AI" cannot.  I believe that 10 years from now (a very conservative estimate given the breakneck speed of this field's advancement), machines will outclass human lawyers just as they outclassed chess and go players (and soon to outclass Starcraft II players).

Facts: Lawyers charge hundreds of dollars per hour, and the adversarial legal system can make the ability of the lawyer matter more than the facts of the case.  I have quoted the pricing of ChatGPT4 in my post, they are significantly lower.

My conclusion: An AI can do a better job as a defence lawyer than most human lawyers for a fraction of the cost.
My prediction: However, given that the legal system is a protection racket, people will never be allowed to take advantages for an AI without paying a lawyer regardless.

Feel free to poke holes in my arguments, I've been known to change my views in light of sufficiently compelling evidence.

[Al Terego]

And my point was that if you're not an LLM (Master of Laws) and you're using an LLM (Large-language model) for legal matters then there is potential for a lot of problems that can be caused by one's ignorance of law and legal procedure, or alternatively one's stupidity with regards to how they believe the law *should* work.

In legal matters, I am stupid.  But a properly trained AI won't be.

Also the stuff relating to confidentiality, evidentiary protocols, disclosure policies, privacy, security, etc.

All of them are rules that AI can be trained on.

And while your whole point seemed to center upon the "Legislators [banning] to protect their caste", my point of why it wouldn't work stemmed largely from the fact that ChatGPT is not a legal person and has no legal standing. I cannot secure ChatGPT as counsel not because they want to 'secure their caste', but because my counsel has to be a legal person registered to practice law in the state and in that area of law.

That is circular reasoning.

IF ChatGPT were able to be used in the courtroom, you would legally be representing yourself as you do not have counsel representing you.

Sure.

This means that Average Joe with ChatGPT is up against Bigshot Lawyer with ChatGPT in court, or worse yet Bigtime-Election-Year Prosecutor with ChatGPT (because why wouldn't they use that tool too?).

The AI would be the equalizer.

If we face each other in hand-to-hand combat, the bigger, stronger, and better trained person will have a decisive advantage.  Give both of us a machine gun, and that physical advantage goes out the window.

I do not believe that Bigshot Lawyer + AI will do much better than AI unhindered by Bigshot Lawyer anymore than a chess International Master (Elo rating below 2500) aided by the Stockfish engine (Elo rating above 3500) can play noticeably better than Stockfish alone.

This means that Average Joe with ChatGPT is up against Bigshot Lawyer with ChatGPT in court, or worse yet Bigtime-Election-Year You likewise then couldn't appeal the results due to ineffective counsel because you never secured counsel and were representing yourself.

Caveat emptor.

[Keelan]

In legal matters, I am stupid.  But a properly trained AI won't be.


All of them are rules that AI can be trained on.

And how do you know if your AI is properly trained if you know nothing of the law? Do you even know what questions you need to ask to make sure that you're getting the right answers? If it tells you you need to file a motion, but you actually need to submit a writ but had phrased your question improperly, would you realize it?

That is circular reasoning.

No, it is not. If I am arrested and I invoke my right to an attorney - to legal counsel - that means I'm asking for SOMEONE that will advise me on the law and speak on my behalf.

Plants, animals, and objects cannot speak on my behalf as they cannot speak, comprehend language, nor comprehend law.

Computers cannot speak on my behalf as they require human input, are not sentient, and cannot be held liable for any statements they may make, and thus if I request counsel and they bring me a computer, I am my own counsel and I am the one responsible.

ChatGPT or any other LLM requires user input to output answers. It cannot spontaneously tell you to shut the fuck up if you start running your mouth, nor can it answer questions for you without input. At best you get asked questions, you get ChatGPT to produce your response with a 'high degree of accuracy', and then you review and submit it.

You are essentially constructing written self-statements every time you use ChatGPT - thus denying yourself the right to remain silent in the process - and you better pray to GOD that each instance of a statement by you matches up with each other because if it doesn't, well ChatGPT can't be liable because it's just a tool and you had every chance to review the statements before you submit them.

The AI would be the equalizer.

If we face each other in hand-to-hand combat, the bigger, stronger, and better trained person will have a decisive advantage.  Give both of us a machine gun, and that physical advantage goes out the window.

I do not believe that Bigshot Lawyer + AI will do much better than AI unhindered by Bigshot Lawyer anymore than a chess International Master (Elo rating below 2500) aided by the Stockfish engine (Elo rating above 3500) can play noticeably better than Stockfish alone.

No, that very specific physical advantage goes out the window; you take some random dude with no firearm experience and give him a M249 and put him up against a highly-competent 0331 Veteran and something tells me that the guy who has used that machinegun for several tours is probably still going to have the advantage.

(This is why everyone who is a halfway decent firearms advocate will tell you if you have $300, it's better to buy a cheap gun, ammo, and get more practice putting rounds down range than it is buying a nicer gun and being able to do less - maybe even no - training.)

Same with the law; random dude with what is essentially the best legal template generator and legal dictionary up against an experienced lawyer with that same legal template generator and legal dictionary?

Not to mention that the same tool that can be used to fight FOR you can be just as effective at CRUSHING you; a substantial component of law EVERYWHERE is the ability to argue your case. LLMs would be completely mercenary in that regard.

And that's not even getting into the fact that unlike chess - which is a highly controlled, highly systematic, highly regulated and highly controlled environment that ultimately has no room for human error intrinsic to it - even with ChatGPT as a replacement for lawyers you STILL have Judges and juries (and other human components of the legal process) where applicable, whom are still human.

The only way this theoretical 'let's get rid of lawyers' notion works is if you replace the ENTIRE legal system with what would essentially be a fully-automated techno-tyrannical judicator which could be easily programmed with biases (or simply have them in there accidentally), and otherwise would be prone to exploitation by state powers that be.

As much as I find lawyers expensive and some to be quite scummy assholes, I would much rather them than some faceless series of if-then-else equations state-obediently ruling with the compassion of a functional psychopath against anyone the state ultimately decides to legislate and weaponize this 'totally impartial' legal network against.

[Oniya]

I do appreciate that Mr. Koetsier (who does not claim any legal experience) provided the data on which exams the ChatGPT was tested on.  I am also not a lawyer, nor have I stayed in a Holiday Inn Express recently.  I am, however, a keen observer of people, and spent a lot of time watching real lawyers in action.

ChatGPT or any other LLM requires user input to output answers. It cannot spontaneously tell you to shut the fuck up if you start running your mouth, nor can it answer questions for you without input. At best you get asked questions, you get ChatGPT to produce your response with a 'high degree of accuracy', and then you review and submit it.

This is something that I have observed lawyers doing for their client's benefit, particularly when the defendant chooses to take the stand in their own defense.  Our hypothetical AI would have to be able to listen to its client's off-the-cuff testimony and anticipate that the client is entering dangerous territory before it happens.  While an AI trained on legal matters might have the ability to predict legal arguments, would it have the training necessary to - for example - anticipate that its client would invoke some half-baked conspiracy theory on the stand?  A human lawyer would have the experience to discuss testimony with their client before putting them on the stand, and the caution to ask questions that don't lead to 'surprises'.  One of the common bits of advice I've heard is 'don't ask a question that you don't already know the answer to'.

[Keelan]

So, this is relevant I think, though I haven't read it in full:

https://www.reuters.com/technology/plagiarism-machines-hollywood-writers-studios-battle-over-future-ai-2023-05-03/

[Oniya]

When I was younger (so much younger than today), I read Robert Heinlein's 'The Moon is a Harsh Mistress', which includes a supercomputer that becomes sentient.  There's an entire section where the protag is trying to explain to this AI 'what is funny'.  That particular bit comes to mind because it was so difficult to distill comedy into a set of rules (which is something that the writers could possibly use to their advantage.)  If I haven't missed anything, the last time that the WGA went on strike, we ended up with 'reality' TV, since that was pretty low-effort as far as actual plot arcs are concerned.  AI might be able to handle the plotting of your average daytime drama, but can it produce moderately reliable humor? 

"Are two types of jokes. One sort goes on being funny forever. Other sort is funny once.
Second time it's dull. This joke is second sort. Use it once, you're a wit. Use twice, you're a halfwit."
"Geometrical progression?"
"Or worse."

[Al Terego]

@Keelan

It seems that we are arguing different things.  I am not advocating for Dick The Butcher's suggestion, just for the option for a person that chooses to represent themselves to be able to use AI as an aid.

And how do you know if your AI is properly trained if you know nothing of the law?

And how to you know that your lawyer has been properly trained?

Or what if you don't have the means to afford an attorney (like sample source)

No, it is not.

It is.  I am arguing against the established rules, and you are saying that my position is invalid because it is at odds with the established rules.  This is the textbook definition of circular reasoning.

If I am arrested and I invoke my right to an attorney - to legal counsel - that means I'm asking for SOMEONE that will advise me on the law and speak on my behalf.

See above for legal counsel affordability.

Plants, animals, and objects cannot speak on my behalf as they cannot speak, comprehend language, nor comprehend law.

A person can choose to represent themselves.  I argue that the use of computerized aid should be allowed.

ChatGPT or any other LLM requires user input to output answers.

Not anymore, and definitely not 10 years from now.

It cannot spontaneously tell you to shut the fuck up if you start running your mouth

Just add voice recognition.

nor can it answer questions for you without input. At best you get asked questions, you get ChatGPT to produce your response with a 'high degree of accuracy', and then you review and submit it.

Are you basing your argument on the limitations of a single implementation (ChatGPT) at his point in time?  Just take a look at how fast the technology evolved since its inception, and it will evolve even faster now that it got the interest of the open source community.

[...] ChatGPT can't be liable because it's just a tool

On this we are in agreement.  My point is that one should be allowed to use this tool if they wish to.

But why limit the discussion to criminal defence in court?  Why not use software to formulate contracts, write wills, find and cite applicable case law, etc.?
There are many fields that require a lawyer who basically acts like a glorified expert system, with the additional fallibilities stemming from being human.

Several years ago I was involved in collision and was charged under the traffic act.  I represented myself, did my homework, spent time in the law library of the local university researching case law, and ultimately prevailed.  Having a computerized agent to assist me could have saved me quite a bit of time and a some anxiety.

Most lawyers to not argue at murder trials.  Hell, only half a percent of US lawyers are trial lawyers.

No, that very specific physical advantage goes out the window; you take some random dude with no firearm experience and give him a M249 and put him up against a highly-competent 0331 Veteran and something tells me that the guy who has used that machinegun for several tours is probably still going to have the advantage.

Not if the machine gun is fully autonomous.

Same with the law; random dude with what is essentially the best legal template generator and legal dictionary up against an experienced lawyer with that same legal template generator and legal dictionary?

I claim that in 10 years time, the "legal template generator" will be better than a human lawyer.

As much as I find lawyers expensive and some to be quite scummy assholes, I would much rather them than some faceless series of if-then-else equations state-obediently ruling with the compassion of a functional psychopath against anyone the state ultimately decides to legislate and weaponize this 'totally impartial' legal network against.

I did not suggest this.  Let's keep strawmen out of the discussion.

[Al Terego]

So, this is relevant I think, though I haven't read it in full:

https://www.reuters.com/technology/plagiarism-machines-hollywood-writers-studios-battle-over-future-ai-2023-05-03/

They are scared for their jobs, and rightly so.

But let's look at some parts of the article:

The union is also arguing that existing scripts should not be used to train artificial intelligence, which would open the door to intellectual property theft.

“We call it the Nora Ephron problem,” August said, referring to the writer of romantic comedy hits including When Harry Met Sally and You’ve Got Mail. “One can imagine a studio training an AI on all of Nora Ephron’s scripts, and having it write a comedy in her voice. Our proposals would prevent that.”
WGA chief negotiator Ellen Stutzman said some members have another term for AI: “plagiarism machines.”

I would say that the same criteria should be applied to determine if a work is infringing regardless of whether it was produced by a human or otherwise.
What prevents a human writer from reading all of Nora Ephron's scripts (or watching the movies) and then producing a work in a similar voice?

“We have made a reasonable proposal that the company should keep AI out of the business of writing television and movies and not try and replace writers,” she said.

Where were you when manual workers started being replaced with automation?

[Skyguy]

I like the idea of AI art. I think it’s got huge potential to allow a lot of people access to get pretty decent quality art for their ideas that they might not otherwise have been able to get. It’s not on the level of a properly drawn and commissioned piece but if I just need a quick picture for an idea it’s great.

I don’t like the tech bros making the AI. I think it was the CEO of Midjourney was interviewed about the ethical issues with having scrapped millions of images without consent to train his AI and now it was cranking out images that were pretty damn plagiaristic. His response was a decidedly uninspiring blend of “No one told us we couldn’t,” and, “LOL already did it.” The main impression I got was that he’d barely even considered it, much less the implications of having ripped people’s art off to make his AI and now he was selling access to it. Like most of their breed, tech bros, there is zero consideration for the ethical implications of anything and barely any for the legal ramifications. Most of the explanations I hear for the ethics are “wells someone’s gonna do it so it may as well be me getting rich off it,” and the legal, “No’s one’s written a law that specifically makes the legally dubious thing I’m doing illegal so that means it’s totally cool.”

I cannot believe I’m saying this but I honestly hope that Getty eats Stable Diffusion’s face.

[ninjawriter]

Where were you when manual workers started being replaced with automation?

You are literally the only other person I have heard make this argument.  One of the frustrating things, for me, about this discussion is that it has already happened again and again to people in other professions.

[ninjawriter]

I believe the biggest issue regarding AI art is that the arguments that are being used to "defend" the artists' rights don't protect any artists.  There are two arguments I see advanced:

1.  Artists whose works are used for training should be paid.

2.  Styles should be protected as intellectual property.

Both concepts are novel.  Until a year ago, the idea that either of these things should be true for any class of artist wasn't on my radar because all artists train from existing art and styles have never been a protected class.  I'm allowed to read a bunch of Stephen King book and write a book that is stylistically identical to Stephen King!  The general artistic term is pastiche, and while it is usually pejorative, it is unarguably legal.  Or it has been legal. 

Regardless, imagine if those two concepts become legal.  A person might say, "Well, we only want to apply these laws to AIs."  This doesn't even stop AI art, of course, it just means that the people who own the most images have the best AI for a while.  Disney trains its AIs on its database of stories, scripts, comics, books, etc., and its vast art collect - many millions of high-quality pieces of art.

And then they come after everyone else.  Everyone who makes a character LIKE a Disney princess might find that their work is now Disney's work because they trained themselves on Disney art and copied the Disney "style," even if the character isn't particularly like a Disney character.  They will be the people with the resources to "prove" that you "stole" from them.

These arguments do not help artists.  This is why, I think, the big media companies are largely silent.  They're okay with what happens either way.

[Dice]

I believe the biggest issue regarding AI art is that the arguments that are being used to "defend" the artists' rights don't protect any artists.  There are two arguments I see advanced:

1.  Artists whose works are used for training should be paid...

Both concepts are novel.  Until a year ago, the idea that either of these things should be true for any class of artist wasn't on my radar because all artists train from existing art and styles have never been a protected class...

These arguments do not help artists.  This is why, I think, the big media companies are largely silent.  They're okay with what happens either way.

This is, bluntly, wrong. There is a reason why when the company made a version of the program that made music they did not fuck with music companies. The issue is not that our work is not protected and we are not owed for the rights of that which we create, it is that we as a whole are not powerful.

But clearly big media companies have in part had their say because when it came down to it, the folks running this clown show stayed the fuck away from the music giants for a reason.

[ninjawriter]

This is, bluntly, wrong. There is a reason why when the company made a version of the program that made music they did not fuck with music companies. The issue is not that our work is not protected and we are not owed for the rights of that which we create, it is that we as a whole are not powerful.

But clearly big media companies have in part had their say because when it came down to it, the folks running this clown show stayed the fuck away from the music giants for a reason.

What company is this?  I know several companies making generative music AIs, including Google.  I can think of no company that stopped working on music AI because they were worried about music companies.

[Dice]

What company is this?  I know several companies making generative music AIs, including Google.  I can think of no company that stopped working on music AI because they were worried about music companies.

Not stopped working on it, used only music that was in the public domain or open source. There was a bunch of articles about this when it was first announced. The it was amusing that these devs would walk all over us (Visual artists) but not take on someone like Sony. Seems all those posts got buried by the news that they are getting sued for copyright stuff anyhow... so what the hell do I know anymore?

Anyhow I can not find the articles I am looking for, but there was a big blow up that pissed off a bunch of us (I am a photographer) because the early music datasets were not including copyrighted works where they did not give two shits about scraping all we had. The reasons given was we have no power but the Music Studios have actual muscle. It is all well and good to claim your datasets do not infringe of the rights of people who do not have the money to sue you, it is another to test that against the lawyers of EMI or Sony.

Seems these days though its all hands on deck, lets see who we can tread on to make the future. Honestly though, if they are going to ride roughshod over the Labels, I guess that is their fight to pick. I just hate that I know my photos are 100% in this mess somewhere.

[ninjawriter]

Dice, oh, yeah!  That issue arose because music companies were the first to get their art behind well-designed paywalls and the aggression of music companies in pursuing pirates.  You can't scrape for music because the music is behind various paywalls or protected by established contracts (such as the music found in movies.)  That's the essence of the Getty Images lawsuit, really.  They allege that the watermarks prove that their work is behind a paywall and therefore protected.  I don't know how that'll shake out - probably by removing the images from the datasets, which won't affect them very much - but that's tangential to my point, I think.

For me, the real issue is to what extent are machines allowed to do what humans do.  Humans study art (without paying anyone) and then based on what they study, create new art all the time.  Every artist I know can rattle off a list of their influences, right?  So far, though, the law has looked the other way when machines reproduce human labor after a study of those forms of human labor.  I do not say this gladly.  I despise it.  But it has been permitted since the beginning of the Industrial Revolution without question.

And I think that Disney and the other major media companies are drooling at the possibility that STYLES and INFLUENCES can be protected.  That would give them the tools to go after any competitors.

[Oniya]

The general artistic term is pastiche, and while it is usually pejorative, it is unarguably legal.  Or it has been legal. 

I'm not entirely sure that 'pastiche' is necessarily pejorative.  There have been several fairly famous pastiches.  'The Seven Percent Solution' is not based on any single Doyle story, most of the Bond movies before 'The Living Daylights' had bits and pieces from multiple Fleming novels - TLD itself wasn't even written by Fleming, and the myriad Conan storied that weren't written by Howard.  And I suppose I'm obligated to mention that the 'Fifty Shades' books started out as Twilight fanfic.

[GloomCookie]

I've said it before but I'll say it again. All work is built on top of the work that came before, and it's really an argument about how closely something resembles something else and where that line is blurred.

Ed Sheeran recently won a lawsuit that he didn't copy the rhythm, chords, etc. from "Let's get it on" by Marvin Gaye. The reason that's critical is because 3-note chords on musical instruments can only come in a finite variety before you start to become repetitive.

Patents are only good for about 20 years, meaning if I went out and invented a new type of motor that was even more energy efficient today (May 11) then it would become open to the public on May 11, 2043. Nothing I do short of come up with a new patent changes that.

Fashion, meanwhile, has no copyright. You can watch a fashion show from Milan and copy that style immediately.

The reason art is somehow separate is because of intense lobbying by corporations like Disney that wanted to take public domain works, add their own flair to them, and then lock in that new style to establish a brand. If I wanted to make my own version of Star Wars, I'd have to wait until 2072 in order to do so without calling it War of the Stars, which even that title might get me sued. The only thing protecting art is lobbying by big corporations, because originally, they were limited to 14 years. If I published a book today, and I lived to the US average of about 75, then there's an additional 70 years on top of that, then my book becomes public domain in 2133, which feels absurd to me.

I understand wanting to protect your brand, but copyright I feel is used more as a hammer to beat down new ideas rather than allowing people to come up with new ones. How many more bland and passe Superman movies will we see before the character becomes public domain in 2034? How will such a character be treated once anyone can publish a comic with Superman in it? Or put out a Superman movie?

Quietly waits for someone to scream at her. Again.

[ninjawriter]

The reason art is somehow separate is because of intense lobbying by corporations like Disney that wanted to take public domain works, add their own flair to them, and then lock in that new style to establish a brand. If I wanted to make my own version of Star Wars, I'd have to wait until 2072 in order to do so without calling it War of the Stars, which even that title might get me sued. The only thing protecting art is lobbying by big corporations because originally, they were limited to 14 years. If I published a book today, and I lived to the US average of about 75, then there's an additional 70 years on top of that, then my book becomes public domain in 2133, which feels absurd to me.

Indeed.  This is why I said that media corporations are silent with current artists proposing that using images to train upon and stylistic similarities are protected classes.  Because, in the end, this gives them the power to go after not just people who make Star Wars but also War of the Stars, War of the Distant Stars, War of the Faraway Galaxy, etc., with much greater power.  I think artists who are relying on arguments that people shouldn't train on images and protect styles are making a Devil's deal.

[Dice]

I've said it before but I'll say it again. All work is built on top of the work that came before, and it's really an argument about how closely something resembles something else and where that line is

That is true, yes, but art is the one thihng that is truly of the individual who makes it. If you do not have Oppenheimer you would still end up with the bomb. You might need five people to take his place but the rules of physics don't change. You can throw lesser minds at a problem and still crack it in the end.

Art is no like that. If there is no Waters then there is no The Wall. If there is no Freddy then there is no Bohemian Rhapsody. Even if we walk on the stones laid by those who came before us, we still make, we still build and we still give onto others the chance to build of their own two hands and feel something of the moment we shaped something into reality with our own skills.

My issues with AI art is not that someone is taking from my work, finding inspiration and building their own craft, that would be an honour and a privilege, it is that a lot of the craft and the self development is taken away from the "artist" and it reduces the value of what is made by the person involved. There is a gift in finding yourself in the moment and building something with your brush, pencil, notebook, camera or instrument. These things are not found in promoting the same system over and over like throwing darts at a wall.

Perhaps I am an old man in the 1910's looking at power tools and thinking there is no grace in them. Maybe I am out of touch, but I have already done this once before (Film camera's my father taught me on to Digital I now use) and that felt like a change of the tech but not a distruction of the skill. AI is something else.

Quietly waits for someone to scream at her. Again.

AHHHHHHHHHHHHHHHHHHHHHHHHH!

[Kuroneko]

Could we please try not to confuse a copyright with a patent? They're two different things. Trademarks are another thing. Patents apply to technical inventions. Copyrights apply to artistic creations. Patents only last for 20 years, but copyrights last for far longer, generally the life of author plus 70 years. Comparing the two is apples and oranges.

U.S. Copyright Office
United States Patent and Trademark Office
Copyright Alliance

The article on fashion also points out that there are aspects of fashion that are subject to copyright. Many fashion designers are highly recognizable for the very things listed in the article that can be subject to copyright: graphics on fabric, logos included in designs, and textile designs.

I've said it before. While all artists learn from what came before, I don't know a single artist (including myself) that wants to copy what another artist did. I don't want to draw like Da Vinci, or paint like Van Gogh, Picasso or Miyazaki. We've already had one of each of them in the world, and we don't need more. We need new things, new ideas, not copycats. 

Points out that there are no capital letters in this post, or any of my posts, therefore no screaming.

[Kuroneko]

It also might be useful to know that copyright existed long before big corporations. They can definitely now use it to establish a trademark, but it was initially created to protect the intellectual property rights of individuals.

US Copyright Timeline


Hopefully, Anton Toynikov's work with Stable Attribution will make credit to the original artists behind an AI image a reality.

[ninjawriter]

Stable Attribution shut down without any fanfare.  My experience with it is that it wasn't very good, I should add.  I took a well-known image from a fantasy artist and told Midjourney to base an image on the Frazetta image I used, and fed it into Stable Diffusion, oh, I dunno, at least ten times.  It only ever returned one Frazetta image as a possible source and it was the wrong image.

[GloomCookie]

Could we please try not to confuse a copyright with a patent? They're two different things. Trademarks are another thing. Patents apply to technical inventions. Copyrights apply to artistic creations. Patents only last for 20 years, but copyrights last for far longer, generally the life of author plus 70 years. Comparing the two is apples and oranges.

U.S. Copyright Office
United States Patent and Trademark Office
Copyright Alliance

The article on fashion also points out that there are aspects of fashion that are subject to copyright. Many fashion designers are highly recognizable for the very things listed in the article that can be subject to copyright: graphics on fabric, logos included in designs, and textile designs.

I've said it before. While all artists learn from what came before, I don't know a single artist (including myself) that wants to copy what another artist did. I don't want to draw like Da Vinci, or paint like Van Gogh, Picasso or Miyazaki. We've already had one of each of them in the world, and we don't need more. We need new things, new ideas, not copycats. 

Points out that there are no capital letters in this post, or any of my posts, therefore no screaming.

No, it absolutely is not, because both work on similar principles. While art may be unique in that each artist is slightly different, fundamentals that make up the core of works stem from the works of others just as patents do. The works of the ancient Greek masters inspired the works of Renaissance artists who in turn inspired others. Work does not exist in a vacuum. Picasso would not have created his famous works had he not studied the works that came before.

The same can be said of other artworks such as literature. Chaucer was one of the earliest writers in Old English and created Canterbury Tales, and then you had Thomas Malory's Le Morte d'Arthur, both works that influenced later works as the language and structure of literature evolved. They would go on to inspire Shakespear, who in turn would inspire others throughout history.

All works are built upon someone else's works, even if it's just a stepping stone or them reading and thinking "God this is so bad, I could do better." Stephen King, for example, states that he got a lot of his earliest inspirations with H.P. Lovecraft's The Lurker in the Shadows, stating "I knew that I'd found home when I read that book." He also read William Goding's Lord of the Flies.

Now, having beaten that horse to death, let me trot out the next argument. There are lots of patents covering software development. As anyone who has ever programmed something before, they will tell you that there's multiple ways to achieve the same function using programming, even within the same programming language. For example, how to program the game FizzBuzz.

https://youtu.be/QPZ0pIK_wsc

Now, if you try and patent FizzBuzz as an example, you can't patent the actual code, but instead the results of the program. That is remarkably similar to how copyright works for literature and art, since each programmer has their own unique style and it's all about the final product. So, if you hold the patent for FizzBuzz, and I produce my own unique program but it achieves the same goal, that's patent infringement.

But if I create a piece of art and call it FizzBuzz, and you then produce a piece of art that looks similar to FizzBuzz, I can sue you for copyright infringement.

Can you see where the lines blur?

Oh, and let me share something else that's common in my industry. My work as an engineer is copywritten by the owners of the company I am working for. Starbucks, for example, has all its plans and documents copywritten. Not trademarked, even though it has the Starbucks logo on it, but there is a large block of text on each document that states the drawings are owned by Starbucks and can't be copied without permission. I can't take a detail from Starbucks and copy it onto a new project.

Is it art? I mean, it's all physics and science, so why can't I copy it? Well, parts can and can't be copied. Starbucks doesn't own the symbol for a panelboard or disconnect, so I can copy that. I can copy information like I want to use Bussmann LPN-RK500SP fuses on the next project. So it has to do with the specific configuration of graphics on a sheet of paper that's copywritten.

So if the components themselves don't matter... then maybe the individual parts and pieces don't matter so much, and maybe it's the final, end product that's actually copywritten. Just ask Ed Sheeran, who just got accused of copyright infringement for having the same basic chords on his song as another.

[HeLaFem]

I don't know a single artist (including myself) that wants to copy what another artist did. I don't want to draw like Da Vinci, or paint like Van Gogh, Picasso or Miyazaki. We've already had one of each of them in the world, and we don't need more. We need new things, new ideas, not copycats. 

If we consider "artist" as "creator", then, of course, I agree with you.
But if we consider "artist" as a "profession", then copying (both style and full-fledged paintings) is a whole type of activity that they have been engaged in the past and are doing now. Because by copying a famous author\a successful picture, a person can get a fraction of his fame and money by doing craftsmanship, not art(so, I hope ChatGPT does not deceive me and these two concepts really reflect the same palette of meaning as the original thought). Especially in a place where the popularity of the original author is not represented. Another city, another country.
The same thing is happening now, just in a less explicit manner (and even then, not always. In musical works, at least in my country, there are some couple of complete copies of works \ parts of works from the USA\Europe just with a translation), and such people have been and will be as long as it benefits those who do it. And for so many of their fans, they will be real artists and even creators, while in fact they are a mix of thieves(in a copyright laws) and creators.

Considering that you suggest not to confuse copyright with a patent, I'm sure you understand this perfectly well, much better than me, because these things were created, including because of such people, complicating their lives and taking them beyond the letter of the law. So consider this an invitation to a conversation. :)


As for the main topic (or rather a more general view of AI), for me AI is another round of progress that can only be slowed down but not stopped. During the Industrial Revolution, people lost their jobs, and professions were devalued, but at the same time, new, other in-demand jobs were created. From the point of view of society, this is progress. From the point of view of a resident of that time, it is a tragedy. All we can do is try to delay or smooth out this effect in time, but it will be, and this cannot be avoided.
Already, some Chinese companies (from the rumors that I have seen) are quite starting to use music written by AI, art drawn by AI, while firing their employees who were previously engaged in creativity.

Because, for the most part, it is not the work of people that is important to their consumers but the final picture. And if it can be achieved in more straightforward ways, people will use it. Not only large corporations but also small artists, developers, and companies. And the human eye will not find the differences between an adequately selected work made by AI from one made by a natural person just because AI is trained on arts made by people and, as a result, adopts its pros and cons. One piece of evidence of this is the victory of the photo created by AI at the Sony photo contest.
One of the exciting things here is the oversaturation of consumers with high-quality things and the search for new, exciting styles through, for example, purposeful product corrosion. Plus, an interesting separate category of AI works trained on the works of other AI and a lot of iterations of this (as they did, if memory serves me right,
AI from DeepMind on the starcraft game), as a result, literally the whole idea of the art may change. This can also become one of the extremely unlikely, but possible ways to use AI for creativity. Separation of AI creativity from people's creativity and taking AI underground, which creates images based on people's works. More of a fantastic scenario, of course.

And as for the professions, the natural creators will remain one way or another. There will always be a certain number of people who care about human labor. It will go from the category of a full-fledged profession to the sort of enthusiast. The entry threshold and attitude towards this profession will change dramatically.
However, with this change, my profession will simply die out in its current form because engineering is about technical interaction with technology for the purposes and needs of a particular situation rather than about creative activity in general. ^^"


As for the ethical problem raised here, I would like AI works to consider people's copyright and respect their work and time spent. This discussion is essential, and discussions in society, as a result, will allow us to come to some kind of consensus.
But one of the problems (and dignity) of the our world is the competition of ideas and views, not only within society but also between different societies. As a result, it can easily be a situation where in country A, AI works are limited by ethics and copyright law. In country B, AI works will be significantly less limited and, as a result, create entirely different things. Worse or better is an open question, but this is one of the points of tension and disagreements between countries. It is also worth remembering that such work on discussing AI should go not only within one society but also in cooperation with other societies.

[Al Terego]

I've said it before but I'll say it again. All work is built on top of the work that came before, and it's really an argument about how closely something resembles something else and where that line is blurred.

This!

The only difference between AI training and artist training is scale and efficiency (the quality of the output and the similarity of content is a different question.

The reason art is somehow separate is because of intense lobbying by corporations like Disney that wanted to take public domain works, add their own flair to them, and then lock in that new style to establish a brand. If I wanted to make my own version of Star Wars, I'd have to wait until 2072 in order to do so without calling it War of the Stars, which even that title might get me sued. The only thing protecting art is lobbying by big corporations, because originally, they were limited to 14 years. If I published a book today, and I lived to the US average of about 75, then there's an additional 70 years on top of that, then my book becomes public domain in 2133, which feels absurd to me.

I understand wanting to protect your brand, but copyright I feel is used more as a hammer to beat down new ideas rather than allowing people to come up with new ones.

Copyright was established at the behest of, and for the benefit of the publishers, not the artists.  And the current terms are more than just absurd, it is outright robbing the commons.

IP protection in the US is based on Article I, Section 8, Clause 8 of the constitution, which states that gives Congress the power "To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries." after which the works should enter the public domain.

The original copyright term was 14 years (plus an option for another 14 years extension) and it required registration.  In my opinion it struck a good balance, which the current regime does not.
I fail to see how life+70 "promotes the progress of science and useful arts" and how it benefits the public at large.  If anything, it removes an incentive for successful authors to continue creating, and prevents others from improving on those works.

Health professional, first responders, and all other "essential workers" are paid for work done.  A heart surgeon does not expect the person they saved to continue paying in perpetuity for the privilege of using their fixed/transplanted/artificial heart, do they?  And yet, the idea that "creative work" should continue paying the bills for one's great great great grandchildren is pushed as sacrosanct.  I guess JRRT should be publishing another Middle Earth masterpiece any time now, and the fact that the law forbids me from singing "happy birthday to you" in a restaurant makes all our lives so much better.

Sophocles said that immoral laws should not be followed.

Quietly waits for someone to scream at her. Again.

I shall scream my agreement.

That is true, yes, but art is the one thihng that is truly of the individual who makes it. If you do not have Oppenheimer you would still end up with the bomb. You might need five people to take his place but the rules of physics don't change. You can throw lesser minds at a problem and still crack it in the end.

Art is no like that. If there is no Waters then there is no The Wall. If there is no Freddy then there is no Bohemian Rhapsody. Even if we walk on the stones laid by those who came before us, we still make, we still build and we still give onto others the chance to build of their own two hands and feel something of the moment we shaped something into reality with our own skills.

If there was no Waters or Freddy, someone else would have risen to fill that void.  They might not have created the same (or even similar) work, but it could have been just as good, or even better.  Nature abhors a vacuum.

My issues with AI art is not that someone is taking from my work, finding inspiration and building their own craft, that would be an honour and a privilege, it is that a lot of the craft and the self development is taken away from the "artist" and it reduces the value of what is made by the person involved.

It is not up to you to decide.  As it was not up to Oppenheimer to make a decision on Hiroshima and Nagasaki.

Could we please try not to confuse a copyright with a patent? They're two different things. Trademarks are another thing. Patents apply to technical inventions. Copyrights apply to artistic creations. Patents only last for 20 years, but copyrights last for far longer, generally the life of author plus 70 years. Comparing the two is apples and oranges.

The ideas behind them and their justification are the same, the only differences are the details of implementation.
See the relevant passage from the US constitution that I quoted above (I use the US model because information on it is easily accessible and I am too lazy to coduct global research for the sake of a post).

It also might be useful to know that copyright existed long before big corporations.

Say what?[/quote]

[...] it was initially created to protect the intellectual property rights of individuals.

It created those Imaginary Property rights.  Physical property rights are natural, without them society will colapse.  There is a reason that we find laws respecting them on clay tablets from over 4000 years ago.

[Al Terego]

Damn, screwed up the BBCode and cannot edit.  Oh well, the link works.

[Kuroneko]

Stable Attribution shut down without any fanfare.  My experience with it is that it wasn't very good, I should add.  I took a well-known image from a fantasy artist and told Midjourney to base an image on the Frazetta image I used, and fed it into Stable Diffusion, oh, I dunno, at least ten times.  It only ever returned one Frazetta image as a possible source and it was the wrong image.

Ah well, hopes for reliable attribution of sourced works through this method dashed, lol. I think the point of SA was to identify the artist, but hopefully new options will develop.

No, it absolutely is not, because both work on similar principles. While art may be unique in that each artist is slightly different, fundamentals that make up the core of works stem from the works of others just as patents do. The works of the ancient Greek masters inspired the works of Renaissance artists who in turn inspired others. Work does not exist in a vacuum. Picasso would not have created his famous works had he not studied the works that came before.

Bolding mine.

Of course work doesn't exist in a vacuum. But the key concept here in the way that work from previous artists influenced others that come after them, which is embedded in your very argument, is that it inspires others. It doesn't inspire them to simply copy what has come before. It inspires them to innovate. Why would I want to do what someone else has done? Stephen King may have found a home in Lovecraft's work, but he didn't copy it. His work is entirely different.

Copyrights and patents are inherently different or the laws regarding them would be the same. Similar does not equal the same. They apply to completely different types of ideas and objects. I've provided the links to the laws regarding the differences. In the end, it doesn't matter what you or I think about it; wiser people with more knowledge of the laws have made that decision, and we have to follow them. I'm pretty damn thankful that my books are covered by copyright. However, they're not inventions that require a patent.

And yes, the Starbuck's mermaid is indeed a trademark - Starbuck's Trademarks. Several versions of it have been trademarked.

Thankfully, we don't have to agree on this issue.

Al Terego, the Romans could never have imagined corporations are they are now. And since the link I posted specifically referenced copyright law in the U.S., it's not really applicable. Again, apples and origins.

Copyright was established at the behest of, and for the benefit of the publishers, not the artists.  And the current terms are more than just absurd, it is outright robbing the commons.

I don't understand what 'robbing the commons' even mean? Artists are under no obligation to place work in the hands of the public or the public domain.

Copyright absolutely benefits artists. My publisher doesn't own the rights to the content of my books. They currently own the right to print ,publish , and distribute them because that's the contract I signed with them. In return, they protect them from copyright violation and I earn royalties.

IP protection in the US is based on Article I, Section 8, Clause 8 of the constitution, which states that gives Congress the power "To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries." after which the works should enter the public domain.

And that term was set by the government. Whether anyone sees the value of the current copyright term or not, it's the law. If you want it changed, then work for change. 

I'm going to bow out of the conversation now. I'm tired of the argument that theft = the democratization of art. Ya'all have fun with it.

[ninjawriter]

The only difference between AI training and artist training is scale and efficiency (the quality of the output and the similarity of content is a different question.

Also that AIs aren't people.  Which is the big objection... though maybe if visual artists seemed to care about other people whose jobs have been destroyed by mechanization they'd have a bit more of a moral leg to stand upon.  But two wrongs don't make a right.  The consequences of AIs destroying human labor are serious and lasting.

[ninjawriter]

Ah well, hopes for reliable attribution of sourced works through this method dashed, lol. I think the point of SA was to identify the artist, but hopefully new options will develop.

Unlikely.  AI developers prioritize developing AI over the systems to analyze AI... but to analyze an AI, you need a system more powerful than the AI.  This means that we can, at best, substantially analyze previous versions of AI.  And, honestly, by this time, the large AI art generators aren't relying very much on extant instances of YOUR art.  Since the art that they generate goes into the database for further training, the large majority of images, by now, are those generated by AI.  Tracing the provenance of any specific image will only grow more complex as the size of the databases grows and the number of variables grows.'

Regardless, it's still a trap.  The people such an analysis will tend to benefit are large media firms because they both have the largest share of art that people are inspired by and they have the most resources to take legal action.  The problem isn't AI scraping the Internet for art.  It's that we live in a society run by corporations.  Midjourney is not your main enemy.  Disney is.

[Al Terego]

What I don't understand is the claim that art AI is plagiarizing artists by training on their images.

The AI is trained on very large data sets.  The percentage of images made by any single artist is minuscule.  Literally a drop in a bucket.  If you ask it to create an image of some scene, it will do so based on the aggregate of the weights the training data supplied.  Remove one artist's images from the equation and the end result will not be much different.

Now if you ask it to create the image "in the style of Albert Oehlen", it's a completely different situation.
But then, if I happen to have a very talented artist friend and ask them to do that, and they come up with something extremely close to the original, is there a difference?

[TheGlyphstone]

I'm not sure theft stops being theft as a matter of percentages. If an AI scraper steals 1 images from 10,000 artists, it's stolen from 10,000 artists. Each of those 1 images being 1% of their total creations isn't really relevant to whether it's theft or not.

[Al Terego]

1. There is absolutely no "theft" involved.  Please use the correct terms.
2. Looking at an image, memorizing it, and learning from it is not "stealing".  Please use the correct terms.
3. Copyright infringement is *absolutely* a matter of percentages.  An image that was 0.1% inspired by your art -- even of those 0.1% were directly lifted from your art -- is not infringing.

[TheGlyphstone]

Needing to fall back on semantic pedantry is rarely a good sign regarding the solidity of your argument. Just going to point that out.

[Al Terego]

Needing to fall back on emotionally loaded words and hyperbole is rarely a good sign regarding the solidity of your argument. Just going to point that out.

[TheGlyphstone]

'I know you are but what am I'? Really? That's literally the argument level of preteen school children. I can't offer any sort of coherent reply to that without burning IQ points, so a winner is you?

[Dice]

1. There is absolutely no "theft" involved.  Please use the correct terms.
2. Looking at an image, memorizing it, and learning from it is not "stealing".  Please use the correct terms.
3. Copyright infringement is *absolutely* a matter of percentages.  An image that was 0.1% inspired by your art -- even of those 0.1% were directly lifted from your art -- is not infringing.

OK dude, I'm mostly ignoring you, but you should read the thread your posting in some. This here puts to rest most of your arguments imo. Also I wish to point out that I believe you to be arguing in bad faith and will not seek to engage with you while you are doing so. Nothing personal mate, just not worth the blood pressure rise to explain how theft works only to have a semantic fight over already established principles.

This is likely to end up being a major issue for groups that have AI art programs. If your TOS claim they own the copyright to the "Art" they make they are in for hell over this.

https://petapixel.com/2023/02/02/ai-image-generators-can-exactly-replicate-copyrighted-photos/

TLDR:
You can get AI programs to spit out a near 1 to 1 of the images they were trained on. So close they would be in legal shit if someone tried to use them.

[Oniya]

As this is my thread, and the conversation has devolved to this level, I'm locking it.  Nobody wins.