Creative Commons license icon

Fur Affinity bans artworks generated by artificial intelligence programs

Your rating: None Average: 3.5 (8 votes)

On September 5th, Fur Affinity announced it will not be accepting art works that are generated by artificial intelligence programs. Recently computer algorithm generated artwork has come under much discussion as machine learning applications such as DALL-E have shown the ability to take obscure prompts and create art pieces based on them.

Content created by artificial intelligence is not allowed on Fur Affinity.

AI and machine learning applications (DALL-E, Craiyon) sample other artists' work to create content. That content generated can reference hundreds, even thousands of pieces of work from other artists to create derivative images.

Our goal is to support artists and their content. We don’t believe it’s in our community’s best interests to allow AI generated content on the site.

Dragoneer indicated in a Twitter statement on how blatant these programs are in assimilating the pieces of other artists as their own. He pointed to an example of a signature placed on a piece giving evidence of those that the program sampled from.

We made the decision to ban AI generated content on FA. I know a lot of people are asking, "but what about..."

Look, the AI-generated "art" is openly copying the signatures of the artists and teams it's sampling from. Maybe even from somebody you know. It's too much. - Dragoneer

Comments

Your rating: None Average: 2.5 (4 votes)

Good for them. The current level of so-called AI art isn't art.

Your rating: None Average: 3.7 (3 votes)

And there was never any bad art on Fur Affinity again!

(I'm all for this rule, too, but you did walk right into that one.)

Your rating: None Average: 3 (2 votes)

The deeper issue is no level of AI art will ever be free of how this kind of AI art functions, as it is simply melding together existing images.

There's a reason why actual copyright officials have been refusing to touch the many eager AI image submissions which have already been submitted. It is an intractable problem.

Your rating: None Average: 2.8 (4 votes)

I don't know. This decision doesn't sit entirely well with me. I agree that a flood of AI generated images are not necessarily good for anyone but AI algorithms are just tools. In a sense it feels like the same sort of snobbishness that might lead someone to say that digital art isn't real art because there's an undo button and tools like blurs and filters which function automatically. That's all true, it's not the same as hand-drawn art but it takes skills to use those tools effectively. When it comes to using AI art, I imagine the better artists will be those that can envision more interesting scenes and write better prompts to generate those images.

As for the algorithms analysing thousands (I'd guess it's closer to millions) of artworks, so? That's what artists do too. You're supposed to look at art from different artists and learn from it. Sometimes that means studying how the poses work, how to lay out the individual aspects of a piece or even copying a style or technique. That's how art movements are born; people liking and copying a particular way of drawing or painting. Whether that's pointillism or chibi or so on. It's how the way we do things changes. Look at the dolly zoom from Vertigo or the circling camera around a frozen scene from The Matrix. Those have been copied, adapted and reused over and over. The AI is not doing anything that humans haven't been doing already.

The fact that some AI art includes a signature is not evidence that it's somehow evil, just that it doesn't understand what it's doing. It has no concept of a signature or fox or anything else. It knows there are words associated with certain features and it tries to figure out what your new words will look like based on the examples it's seen before. People do the same thing; they will copy the behaviour of others around them, even if they don't understand the reason why people are doing that behaviour. If anything this makes the AI a little less artificial and a little more human.

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None

Now if an AI is smart enough to know to obfuscate its inspirations of its own pieces, or grasp the concept of copy rights, then we'll know that Skynet is not too far behind.

If Skynet were Disney...

Your rating: None

Or it could be right under our noses... https://www.skynet.net/

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None Average: 2.7 (3 votes)

I would rather not make an account just to rebuttal this. But the problem with AI-Generated content is that it's not the same as a human practicing for years, studying their favorite artists from how they start the process; from how they sketch to how they draw anatomically and to the end result. Keep "end result" and "the process" in mind as you're reading this.

For this, we're going to create two people. One is a Japanese artist, let's say he goes by the name Kojiro on the internet; and the other is an American artist called FluffyFoxLover. Both are entirely fictional and my bad if I accidentally just named two popular artists. (But I mean, come on. FluffyFoxLover? That's a really obvious name to go by.) Kojiro has been practicing art for several years now, and is quite popular on Twitter. FluffyFoxLover likes the kind of work Kojiro makes, and decides to study his art for a day to get an idea of how he draws furries. They take a bunch of Kojiro's works, and uses basic shapes to understand any patterns on how a character is drawn in his art style. This is how I, and a lot of other artists learn to draw a certain way.

Now let's create another fictional character, let's call him DogeCryptoBro1337, Doge/Crypt for short. And let's say he paid a lot of money to get access to some of these high-quality image generators that aren't Craiyon. Crypt sees Kojiro's work, and thinks, "Hey, I want my fursona drawn in that art style. Maybe the AI that I paid early access to will be able to make something out of it." And so he punches in a few descriptive words that will end up creating said image. This time, it took the AI about 5 minutes or less to produce images that fit the description. All it did was look at images, found patterns, copied those patterns, and did none of the hours of work that FluffyFoxLover had to do.

There's a stark difference between the two. FluffyFoxLover took time out of their day to understand how Kojiro drew, and all DogeCryptoBro1337 did was tell a computer to do something. I was a bit offended by how you said it would lead to the same snobbishness that traditional people said digital art isn't real art; but is it really the same as two people fighting over what medium they use to make art compared to someone asking an AI to take thousands of artworks (likely sampling from artists that did the same thing that FluffyFoxLover and Kojiro did) and copy only the patterns it sees to fit the criteria of (what can only be assumed as) an essay to describe an image that Crypt wanted?

Art is not about the destination, it's about the journey to get there. I don't follow the idea of separating an artist from their art, because there is barely any distinction from the two. There is a lot of work that goes into Kojiro's work, and he too studies his favorite artists, or even what is taught in schools, the same way FluffyFoxLover has to.

Yes, technically and fundamentally the idea of artists looking at what other artists do and learn is close to an AI being trained to draw based on different images. But it's only if said AI will also: start with a sketch, understand anatomy and can draw different body types, species, hair styles, etc. Then it proceeds with either inking and then flat colors, or for digital painting proceed with the coloring and then rendering for both. Then yes, it's fundamentally the same. But taking shapes, patterns, and taking the images as well so they can be stitched together into something else is not the same as the process that non-AI-Generated art goes through. Their end result is the same, but ultimately the art being made by an artist will always have more substance to it because there was more effort put into it. (If you want to argue that there was also effort put into programming the AI, remember that in programming your best friends are CTRL + C and CTRL + V. The only effort put into training the AI to create images was keeping the electricity up in the server room. That's it.)

I do not put my faith into AI-Generated Content as being revered by artists, because it really is taking away their jobs. Their whole life's work even. I'm sure you would be offended too if say you've been working as an artist, doing commissions to survive for years. And then all of a sudden some stuck-up nerd in America creates an AI that's perfected your art style and can create images that you would need hours upon hours of work to even make.

Your rating: None Average: 3.7 (3 votes)

Interesting debate, though I want to say a few things.
My main issue with some of that one moral debate is that I feel like one is behaving that if something innovative happened and it effected certain traditional jobs in a negative way, then that's considered a problem because a lot more people uses a lot of certain innovations instead. It's like with cars becoming a thing, effecting a lot of older traditional jobs related to horses. The issue is that nobody was ever morally obligated to pay for artists in the first place alone just as much as nobody was morally obligated to pay for horses for main travel, which is why I scratch my head over this issue.

Another thing I scratch my head over is the artist's "style" debate. Nobody could ever own it in the first place if I'm assuming right, and if so, to bring up art style now (as if it's an artist's "right" when it never was) due to the effects of a machine is essentially the same as trying to argue that we can't copy ideas of certain foods (e.g. off-brand coke like soda) because some people buy those instead of the main brand's version.
If traditional artist A copied another traditional artist B's style and such traditional artist somehow negativity effected the original artist, did traditional artist A did something wrong by copying the style, even though I heard you can't copyright an art style? The effects of certain AI generating art with copying certain art styles has the same logic other than it being a machine instead of a person.

Don't get me wrong, I do hope many good artists will keep their job. it's just hard to bring morality sake in it due to some issues, even though I can understand where some are coming from. Though there might be some Copyright risks and/or some other risks. Maybe there are already ways law can block certain AI machine usages. As for the idea that some are losing jobs already because of AI, I am not too sure if there is evidence of that at the moment and I feel like there are some artists who actually wants to use good AI art results, but not always.

Your rating: None Average: 2.3 (3 votes)

Yeah, it's still just a conservative, anti-innovation argument. It's no different than some arguments you could have had against electric lights (Taking away the jobs of those that used to light the lamps) or the arguments we see about moving away from coal (The miners will lose their jobs). Those are real problems with a transition but they are just an argument against progress.

If one wanted to make an argument against AI, or more accurately for artists, then it should be along the lines of "here's something artists offer than AI does not." If the AI can do exactly what artist's do but for cheaper, then we should move on like we have with every advance in history. If artists can offer something that AI can't then there's no problem for artists as they can focus on that unique aspect. Perhaps that unique aspect will be physical artwork which still needs a person to create it. Perhaps it will just be the prestige; like you can buy manufactured furniture from IKEA or you can get unique furniture from a carpenter.

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None Average: 1 (2 votes)

Yeah, but no lamp lighter/coal miner was arguing that what they were doing for a job was art. The thing artists offer that an AI does not is art. Like, the unique perspective of the artist is what makes art art.

I am literally having to explain why you should protect people's unique creative voices to the fucking free speech advocate. Again.

Your rating: None Average: 3 (2 votes)

The AI is not making these things all on its own. It's following the input from a person. The AI product is the AI assisting a person to express their own unique creative voice. It's not only people that have the skills to draw and paint and model that have creative ideas. Many people have those and if AI offers them the opportunity to make those ideas into images that can be enjoyed, what's wrong with that?

This does not stop artists from doing what they do. And if they are offering a unique perspective that AI can not, then they will still be needed for that unique perspective.

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None Average: 1.7 (3 votes)

Well, I think you've run the opposite direction and are severely overestimating the ability of AI to actually create something approximating an artist's visions. I mean, people are not using this to finally provide the visuals for their comic strip they didn't have the ability to draw; it's basically just meme like jokes of the "hey, AI, make SpongeBob, but directed by Zach Snyder" which is fun like for five minutes then it gets boring. I mean, your definition of what constitutes "speech" worthy of protection oscillates wildly, so maybe that is what you're defending, but, I don't really know.

I mean, I'm going to e621, which has been kind of the last bastion of shit banned everywhere else, and ... I'm getting nothing. People aren't using AI to make art because it's just not that useful. It's kind of like NFTs; non-furries could tell it wasn't actually very useful when furries didn't adopt them because we're like the last group on the planet that has an actual art economy going on, so if we weren't using NFTs, they weren't actually useful to working artists. Likewise, if the website that doesn't and probably won't ever ban them wasn't crap flooded by them when they were banned elsewhere, it probably means nobody was seriously using them to begin with.

The images created just aren't, ironically, manipulable enough to be really useful. I mean, going back to the "finally drawing my comic strip" example, something as simple as a character staying on model from panel to panel is nearly impossible. Maybe this is a question that will become relevant as the technology advances, but right now it's just not really at a point where's its anything other than a gimmick.

Your rating: None Average: 1 (1 vote)

I've got to be honest, I don't quite follow most of what you're trying to say here.

Can you explain where you think the potential free speech issue is? I can see someone arguing that banning AI art is a violation of someone's free speech but I can't see how you think allowing it is somehow a free speech issue. It's not stopping anyone from saying anything or posting whatever art they want.

I'm also not sure what you think AI art is capable of. In your first comment, I thought you were worried that people would use AI art because it was maybe faster, cheaper or easier and that would be a risk to artists' jobs. But in this post you denigrate AI art to such an extent that I'm not sure how you think it could be a threat. If "nobody [is] seriously using them" then the fuss over AI art is a little silly, no?

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None Average: 1.5 (2 votes)

Well, partly, I mean, a way to create "speech" has been banned; that no one is apparently actually interested in using this mode of "speech" should not, in theory, deter a "free speech" advocate such as yourself from, well, advocating for it. But, you know, wild oscillations, so I have no idea if you actually care.

But mostly, this conversation started because your comparison of the completely utilitarian, non-speech activities such as coal mining and lamp lighting to the creation of art is just, what we like to call in technical terms, just fucking awful. What I'm saying is, you don't seem to, historically, actually put much value on speech, especially given your otherwise hardline in general (though wishy-washy in specific) stance on free speech. You don't really seem to respect speech much, for all your whinging about it. And comparing making art to fucking coal mining is, frankly, insulting. Double insulting since you like to lecture everybody about "free speech," which makes it seem like your advocacy is not based on actual appreciation of speech, but an awareness that your own speech is so awful that "you have to let me say it because free speech!" is all you got. So, I had to tweak your nose on this.

Of course, in hindsight, I did barge into a conversation otherwise already in progress, and so on one hand there is some context within the post you were replying to that I may have ignored, but, on the other hand, as I said, the comment was not entirely out of nowhere, either.

As far as AI is concerned, it isn't very interesting as art; it may become more insidious with time, however, so I'm okay with nipping that in the bud.

Your rating: None

Okay, that makes sense. But it also tells me that the arguments you are most upset about are not ones that I am making.

You brought up free speech in the comments, not me. While I see a potential to make a free speech argument in favour of AI art, it's not a point I particularly care about nor does any of this seem particularly relevant to free speech. For me, free speech is about the exchange of ideas (including art) and none of this is about banning any particular ideas or content. The exact same picture would be allowed on FA if drawn by a human but banned from FA if generated by an AI, that means that content is not the important part. As I said before, I see the AI as a tool. Banning art that is generated by AI (with some human input in telling the AI what to do) seems equivalent to restricting whether you allow stories that were written by hand versus a typewriter versus a word processor. It's a discussion about the steps in between the idea and the end product.

You are also very much misinterpreting my comparison. I am in no way intend to compare coal mining or lamp lighting to creating art. My comparison was on jobs that were taken away by the progress of technology. If you would prefer, there are jobs that are related to the exchange of ideas which have gone away. For example, jobs involved in the production of physical newspapers and magazines have disappeared as journalism transitioned to digital. Or other creative industries like carpentry. Most people get mass-produced furniture nowadays rather than individually-made pieces by a skilled carpenter. It hasn't completely died but now only the very rich have hand-made furniture by a master craftsman. Same with tailored clothes. I'm sure that those who still make clothes and furniture by hand and to request will tell you that there is art and expression in those crafts too.

"If all mankind minus one, were of one opinion, and only one person were of the contrary opinion, mankind would be no more justified in silencing that one person, than he, if he had the power, would be justified in silencing mankind."
~John Stuart Mill~

Your rating: None Average: 3 (1 vote)

Perhaps more relevant is whether AI methods are like tracing. That is a method of working which is not considered acceptable in most furry art communities without explicit permission. The content might be fine by itself, but the means of production are not kosher. There is also the same underlying motivation of unfairness - that someone is benefitting from work that is not theirs, to the detriment of the original author who sees their work duplicated by others, rather than getting the commissions themselves.

As for cabinets, it is clear that AI can deliver them, but far less clear that it can deliver the ones people want, if they have something specific in mind. In that light, it may be most successful in taking away the business of those who don't know what they want.

Your rating: None Average: 2.3 (3 votes)

To be fair, I find this decision based off what might be considered a 'moral outrage' and that's what actually bothers me about this ban. If this was more about the fear of potential Copyright infringement and/or other trouble involving law, then I would understand more.

Why ban good pictures that were heavily more creative through generation just because what, it copies an artist's style? Or a bunch of multiple pictures to be used for what might be ruled as transformative? What's the difference between that and what thousands of furry artists has always been doing? Just a machine. But again, if this ban was out of uncertainty of potential legal issues, then I would not be as much bothered.

We seriously need to expand the public domain with new good arts, assuming a lot of these AI arts can be public domain of course. There is the fact that AI-generation programs can sometimes help create new good shapes to be inspired for good works. Does the owner of FA have problems with many arts inspired by a result of one too?

Your rating: None Average: 3 (1 vote)

In the US AI generated art can't be copyrighted so it's all in the public domain.

Your rating: None Average: 1 (1 vote)

I think I've heard of something related, however some pictures might end up being way less transformative. If for example, one took an image of Micky Mouse and another one with glasses and all it did was add glasses on it with the majority of the picture obviously looking like it just merged two Copyrighted images together, then we might have an issue involving Copyright.

Though less related to your comment and more of an added take to my original comment here:
There is also a risk of being less transformative with traditional art too.

Your rating: None Average: 5 (1 vote)

In the UK, however, copyright is assigned to whoever arranged for the work to be created by a program:

Wanting to encourage investment in AI in the 1980s, Parliament created a category of “computer-generated works” in section 9(3) of the Copyright Designs and Patents Act 1988 (CDPA). These are works which are generated by a computer in circumstances where there is no human author. The author is therefore deemed to be the person “by whom the arrangements necessary for the creation of the work are undertaken.” To date [Dec 2019], there has been no case law to answer the question of whether there is any requirement of originality for computer-generated works under section 9(3).

There have been consultations on changing this, but as it stands ‘computer-generated work’ is protected for 50 years from the date on which it was made - shorter than the normal 70 years, but still essentially the same form of protection. The owner is the arranger, or their employer, not the AI.

The UK has also permitted text data mining - if you can legally access it, you can make a model based on it; the licensor cannot carve out an exception for this purpose. It'd be logical, though not automatic, that this approach might be applied to images as well

Your rating: None Average: 1 (4 votes)

Nerdy Diamond Guy, if there was an AI that generated artificial porn from real pictures of naked children, would you use it?

Your rating: None Average: 1 (3 votes)

This is probably the most succinct and funny question in this wasteland of a comment section.

Is Nerdy afraid of answering?

Your rating: None Average: 5 (1 vote)

The BBC did a piece about this. It's an active item of discussion among Inkbunny staff. I'm not sure we'll come to the same conclusion as FA.

I see an opportunity to democratize artistic expression - at least in the area of conceptual art, as opposed to detailed commissions. Imagine all those stories that might be enjoyed as comics, or animations, but realistically will never even get a unique cover image without such tools.

There will always be room for artists to use the tools available to create artwork of a higher order; just as synthesizers and drum machines did not end the music business, but transformed it, opening up new possibilities. But I can foresee a time where there is less need for cookie-cutter digital art, and that may impact some artists, until they adapt.

If nothing else, this may promote a resurgence in traditional art, which I feel has been undervalued within fandom in recent years. Certainly, you can get a machine to duplicate that, too, but it may be more time-consuming and expensive, such that it cannot scale.

It should probably be possible for people to identify and, if desired, avoid such content. Credit is important, too - and yes, it is likely going to be contentious if a system has been trained on a large number of images, but that is far preferable to one which uses a small set of work and clearly duplicates the output of a specific group or person.

From a sheer resources standpoint, there may also be capacity issues with the number or uniqueness of such images generated, but this is already something sites have to deal with when hosting 3D art or photography. Nevertheless, with the tools that exist today, getting what you want beyond the initial prompt can still take a significant amount of time - and this indicates there is a level of artistic discernment going into the process.

Is this a crisis or an opportunity? Or both? It'll be interesting to see the results.

Your rating: None Average: 5 (1 vote)

Interestingly I think this is less of a threat to furry artists then general artists in a way.

Furry art is usually character centric in focus, and not really scene focused.

A.I. can't create a new fursona out of thin air, it would have to be derived from those it can pull off line.

A.I. can't give the characters, character. It can only make images.

From what I can see, its ability to do nuances of characters is a bit rough, it's a bit better at landscaping and general works that are not front and center.

An example here by Rick Griffen shows him putting his own character on a background generated by an AI. Of course it has a cyberpunk theme.

In a way I see that the artists are now facing the same problem as comedians. Comedies go around claiming that comedy is 'dying' but it's more a issue that the delivery of comedy has fundamentally shifted. Stand up can't survive in the internet world where a great bit is basically shared around the world before the comedian can even get on the bus to get to the next city on their tour. The only thing that can survive is the improv style, being able to make up bits in the moment. The days of toiling over a typewriter for a set that you can use for a whole year are over.

In the art side, making single art pieces may erode in favor of highlighting character and content beyond stills. An AI can make a nice 'picture' but it cannot world build off the picture. It can't make a graphic novel. It can't create character or lore. Even if AI masters the ability to make a piece, it cannot create a puzzle.

Now if you like making single pieces, your defense would be similar to those in stand up comedy circuits. Kill the internet. Smartphones are taken at the door. Don't put your single pieces online and starve the AI from the ability to access your content. Go analog.

Furry artists will be fine. It's the ones who are doing still-life or landscape that should truly be more concerned.

Your rating: None Average: 5 (1 vote)

1) So i looked at new version of FA upload Policy and
my main complaint is it is not obvious if a background created by an AI is ok.
Example: "Ronin" (https://www.furaffinity.net/view/48423989/ , "posted Aug 5, 2022")
was made by Rick Griffin
who used "MidJourney" AI (to some extent) for the background.

2) related article at Ars Technica:
https://arstechnica.com/information-technology/2022/09/flooded-with-ai-generated...

Your rating: None

Given this is FA we're talking about, the popularity of the artist does have some metric in rule enforcement.

Post new comment

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <img> <b> <i> <s> <blockquote> <ul> <ol> <li> <table> <tr> <td> <th> <sub> <sup> <object> <embed> <h1> <h2> <h3> <h4> <h5> <h6> <dl> <dt> <dd> <param> <center> <strong> <q> <cite> <code> <em>
  • Lines and paragraphs break automatically.

More information about formatting options

CAPTCHA
This test is to prevent automated spam submissions.
Leave empty.