Artificial Inteligence: why I'll not hashtag my art #HumanArt, #HumanMade or #NoAI

Published on

"Cool, which AI did you use for this?", "his work is definitely AI", "this is AI art and I find it disheartening"... Here is a sample of the comments I am getting more and more about my art.

And this is not fine.

In a world where now legions of AIartists are invading platforms like Social Medias, DeviantArt or ArtStation, I see the opinion of the crowd starting to put digital art and AIArt in the same bag. As a digital artist who creates my artwork like a real painting, I find this situation very unfair. I'm using a tablet, layers, digital paints and digital brushes. I put hours of hard work into it. I don't just type in a prompt and hit enter to get my images.

That's why I started hashtagging my art on social media with #HumanArt, then #HumanMade as a clear indicator that my art was 'handmade' and not using Stable Diffusion, Dall-E, Midjourney or whatever AITools are out there now. I wanted to disambiguate that to stop getting the kind of comments I quoted at the beginning of this intro. But which hashtag is the best for that?

I had no idea, so I started a poll on my Mastodon timeline:


The Results

With 954 people voting (thank you), #HumanMade got 55% of the vote compared to 30% for #HumanArt.

But what changed my mind was the diversity and richness of opinions I received in the comments. A lot of them were unlisted, not public or in PM so you can't read them. But they have definitely changed my mind on the subject. That's why I decided to write this blog post and talk about it a bit.

Critiques about #HumanMade and #HumanArt

Firstly, #HumanArt felt like a contradiction to the famous #FurryArt tag for the furry community. Good point, and I don't want that.

Then #HumanMade was criticised because the AI was also human-made, so it lost its point. But most people could clearly understand what #HumanMade would mean under an artwork. So it didn't steal its 55% of the vote.

I also had a lot of #HandCrafted, #HandMade, #Art and other suggestions for variations.

The popular #NoAI

I also got a lot of suggestions for #NoAI hashtags. I got a lot of funnier, and mostly more salty, variations. It was a lot of fun, but I don't want to rule out all artificial intelligence. Some of them might be based on ethical datasets in the future and be good tools. I'll explain that later in this post.

In any case, I have always tried to be in a mindset of 'for something' rather than 'against something'.

Ai artist should be the one who tag their post

This was also a very popular feedback in the comments. Unfortunately, AIArtists rarely tag their work as seen on Social Medias, DeviantArt or ArtStation. And I understand them, because they have too many advantages not to.

First, they can fake an artist identity with little effort. Then they can give their art something more legitimate in their eyes and, by extension, in the eyes of their audience. Finally, they can probably avoid a lot of hateful comments and reports from anti-Ai artists on the platform.

So I don't expect them to ever do it. I hate this because it is unfair.

But lately I started to enjoy this behaviour in a new way, because this faking could be what wastes all the data sets and training models: the AI eats itself.

Don't Hashtag at all

The last suggestion I often received was not to hashtag at all.

This was because writing #HumanArt, #HumanMade or #NoAI would immediately mark the post and the art as premium content for training future datasets. As I wrote in the previous chapter, getting human-made datasets is the future challenge for AI. I don't want to make it easier for them.

I can still write my ethics about "not using an AI image generator based on unethical datasets" in the info section of my social media profile, or just link to an article like the one I'm writing now.

Conclusion and thoughts about AI

So I have made my decision: I will no longer hashtag my art with #HumanArt, #HumanMade or #NoAI.

I will continue to publish my digital art online, as I've been doing since the early 2000s. I will continue to publish all under a permissive Creative Commons licence and with the source files, because that's how I like my art: free and libre.

Unfortunately, I'll never be able to prevent unethical AI companies from siphoning off my art collections. The damage has already been done: hundreds, if not thousands, of my illustrations and comics have been used to train all their AI. It's not hard to find evidence of this (e.g. on or on the browser for the Laion5B commons training dataset).

I don't agree to that.

What are my options? Not much... I can't remove my art one by one from their database. My art has been copied to many wallpaper sites, galleries, forums, other projects. I don't have the resources to do that. I also can't exclude my future art from all the next scans. Also, protection methods like Glaze seem like a really pale solution to the problem, I'm not convinced by it. Same goes for heavily watermarking my art...

Don't get me wrong: I have nothing against AI technology itself. It's everywhere these days. In phones to enhance photos, on 3D software GPU denoisers, on translation tools, behind search engines, etc... The technique of neural networks, machine learning on datasets, is proving to be very efficient for certain tasks.

Even FLOSS projects like GMIC are developing their own neural network library. Of course, they'll be based on ethical datasets. As usual, my problem is not with the technology. It's with the governance and ethics of the one using such a technology.

For my part, I'll continue not to use generative AI in my work (Stable Diffusion, Dall-E, Midjourney and co). I have experimented with it on social media in the past, sometimes seriously, sometimes impressed, but mostly sarcastically.

I just don't like the AI process...
When I create a new piece of art, I don't express the idea to myself with words.
When I create a new piece of art, I don't text the idea to my brain.

It's a more complex mix of emotions, shapes, colours and textures. It's like capturing a rare scene from a dream that temporarily visits my brain. It doesn't require a translation layer of words. When I do this, I share an intimate part of my inner dream world. It goes beyond words to reach certain emotions, nostalgia and feelings.

With AI, AIArtists simply type in keywords for a subject. They season it with keywords, target the imitation of an artist or a style. And then they let randomness produce a result for them. Then they discover a result with - of course - embedded pictorial emotions, shapes, colours and textures. But are the emotions their own or just a by-product of their process? Either way, they can own those emotions.

AIArtists are just diggers for cool auto-generated artworks, the new digital Readymade of our time. This technology seeks productivity for less effort and less cost. I think it is very appropriate for our times. It provides many writers with cheap illustrations for their book covers, editors for their articles, musicians for their album covers and AI artists with portfolios...

I understand we can't go back, this audience is now addicted to this empowerment. They can finally get illustrations quickly and cheaply. They'll treat any artists who fight against this as Luddites...

But I'm going to go on here and say that I personally dislike this art, because it tells me nothing about its creators. What they are thinking, what their aesthetic taste is, what they had in them to draw that line or that brushstroke, what light shines into them, how they cover up their mistakes, their delicious inaccuracies with make-up... I want to see all this and follow the life of people, piece by piece.

I hope you will continue to follow and support my art, comics, articles and tutorials for the same reasons.

Header ilustration: "This is not fine", sources and hi-resolution here − CC-BY 4.0


link Fernando  

Perhaps the best way to troll the system is using the hashtag #maybeAI :P

At first I was very worried about AI but it is as you have said, nobody can replace an artist's spark

link rv  

Hi, David, and thanks for your time and work. This seems to me like a really deep, informed and interesting piece of analysis and opinion, even without considering your final position - which I find very good.
Hope it gets the attention and diffusion it deserves.
Best regards and good luck with everything!

link J  

Dear David,

As an artist, I have thought long and hard about this. I have concluded that to protect ourselves we must integrate AI into the fabric of a platform purely dedicated to 100% human-made art of all forms: poetry, writing, dance, art, music, songwriting, video, and yes even graphic design. Now, you may be asking how one proves that art has not been generated by AI. Well, that is where the AI comes in. The built-in AI would serve to improve the ability of the algorithm to pick up on the use of AI in the generation of the art posted on the forum/website. Of course, this would have to be a well-funded non-profit organization/foundation in the sense that AI would be revolutionary in the field of detecting AI and plagiarism. As we have already seen, companies are citing the use of AI in their advertisements. They may say 50% of this advertisement was generated by AI. Soon, I believe anyone who uses AI for purposes that involve monetary gain and influencing the public will be federally regulated - potentially on a universal scale. And I believe companies will be enforced to be completely transparent and heavily regulated when it comes to the use of AI for any monetary gain. I think you are a forward thinker and the use of hashtags will not only deem your art as a soulful creation but humans will crave a REAL humane use of art, and the ability to tell the difference through a well-funded algorithm that will protect our human right to express ourselves while protecting what makes us truly human: Art. All that being said, I would love to further this discussion in the community of engineers, ethical AI researchers, and all of the artists who will be affected by the rise of this technology that is indeed here to stay. What better way to stay ahead of the game than to foresee the future? I don't see anything wrong with using the hashtags #realart #protectart #realartist #genuineartisty #genuinearts #pureart #authenticart for I truly believe this will only keep you ahead of the game. Being proactive and staying hopeful is all we can do. Let's bring this vision alive together and form A platform dedicated to all of us humans who want to share our art and protect our human connection to art. Let me know if there is a safe way for us to talk. Thanks again for taking the time to read. We got this!

link Amelia Syreth  

People are now awakening to the fact that we live in a culture of industry, art being an industry in itself now. The problem never has been Artificial Intelligence but rather the way most people produce and consume art that enabled it.

On one end of the scale: the 'modern' art which have obfuscated its alleged depth behind an aura of repulsiveness and abstraction. On the other: beautiful but ultimately shallow illustrations that at best tell us about certain interests the artist might have but little else. Humans have filled countless LED-laden bare white cubes with the former, but now AI has given us the ability to maximize the latter. That does not mean everyone will flock to that side of the pond now.

Rather I think people will develop awareness – and consequently habits – to accomodate for passive and active ways to consume art, and look for balance between the two: art that is beautiful but also has meaning that makes you think about it. Oh, certainly we will not cease thoughtlessly looking at pretty pictures of fanarts and wallpapers for our computers, and certainly much like with fast food, people will pay en mass to do so, but I also think more people will (and already are in a sense) want to seek out more beyond that now without falling off the 'deep' end about it. I said from the beginning that those whose works are followed by ones who stick around for the artist behind the artwork don't need to be scared.

Chances are they were always the one who struck the balance.

link JaK  

Stay strong, David! Real artist like you have all my support!

link Donkey Kong  

I can't feel dismayed by AI art. I've played with it myself. Its fun. I enjoy making digital art on a graphics tablet and I don't plan to stop, but AI is very interesting. I can see myself using it for art elements like 2D or 3D textures, I enjoy it. I don't want to brag about being a Hashtag human artist or whatever, because I can see myself encorperating AI into my work. Not a replacement, but another tool in the toolbelt.

I think "ethical datasets" are capitulating to rent-seekers. Images are discarded after training neural nets, and cannot reproduce the exact works they're trained on. There is nothing unethical or illegal about using copyrighted images in datasets. LAION and Stable Diffusion are being punished for their transparency, for being FOSS with a somewhat open dataset in LAION. Midjourney and Dall-E are proprietary software, with no transparency on data imputs, and avoid criticism by being too opaque to criticise. Yet, the LAION dataset is hand-curated for SD training data, you can't tell if your work is in SD even if its in LAION.

You should be sceptical of Glaze. AI users aren't bothered by it at all. It's barely a discussion. Any protection Glaze supposed provides are removed during the standard cleanup dataset images receive. Glaze is also an AI art program itself, its built on Stable Diffusion! If you take a principled stance against AI art, you wouldn't use Glaze. The proliferation of Glaze shows people care more about rent-seeking than a principled anti-AI stance.

link Lilly Mist  

Thank you for this! Just keep up the good work and have fun with it 😀 No ai will ever replace the real artist.

link Sylvain  

Hi David,

Pardon de m'exprimer en français mais je n'ai pas le vocabulaire pour la suite ^^'

Je me demandais, puisque tes dessins sont sous licence Creative Commons, est ce que cela ne veut pas dire que les dessins créé à partir des tiens ne doivent pas être eux aussi sous cette licence pour exister ? Et dans ce cas, tu pourrais potentiellement te retourner contre les sociétés qui ont développer les ia car elles ne permettent pas de respecter la licence de tes œuvres ?

Sinon je trouves ta démarche excellente, merci pour tout :D

link Varflock  

AI "art" managed to annoy me in a few ways already. I find it hard to search for art on Instagram for example by using tags without stumbling upon lots of images that seem interesting at first and turn out to be a product of AI if you look closer. I don't like it, because I want to see someone's genuine work and insight in creating something inspiring, not a machine attempting to simulate the end product of that. In a way art is like sport - we care about people working hard to jump over a tall obstacle or to run really fast, but if someone shows us a robot jumping even higher or running super fast, we won't be so interested in that (well, the military probably will, but for a different reason).

Now imagine a situation where some people buy extremely efficient robots to represent them in sports and try to make them compete with actual people. They could talk about how it makes things more fair and equal because everyone can participate in sports now (like AI "artists" promote generating images), but we'd still know it's not the same at all. We respect both sport people and artists for their effort and skills they honed for a long time. There's nothing worth respecting in generating a similar end result with no work.

A few weeks ago I saw an "artist" promoting her comic art clearly generated by AI and pretending she painted it. People confronted her about it but she kept lying, saying how offended she is by such accusations and calling others jealous even though it was easy to tell that her comic panels look a bit off, clearly change styles all the time and have strange issues like missing fingers on characters' hands mysteriously regrowing a moment later. She also refused to show any art in progress or sketches. That's another thing I'm worried about - cheaters pretending they painted something themselves even though they only wrote a sentence and clicked "generate". It's still easy to tell if they do that in most cases, but AI will only get better at simulating art over time.

I'm sorry for writing such a long comment here, but I wanted to share my opinion.

link Cliff Pratt  

It is unfair when people claim that your art is AI, but what do they mean by that? That you've put in a bunch of words and out popped your art? That is nonsense. That your work looks like it was generated by a computer? That's insulting. So I would suggest that you just ignore stuff like that.

But. I bet your tools already have AI features in them. When you use a fudge tool. When you use a tool to sharpen a line. Someone had to code the intelligence behind these tools. When you soften an area of your image someone worked out how to do this and built the intelligence into the program. This where AI will shine in the future. It won't be necessary to use words to control the AI of the future. You will just draw a line or shade an area with colour, and the AI behind your program will help. If you don't like the help, you will just use your brush to change things and the AI will remember it for next time. The next time you draw a new scene the AI will remember the shadings and the fudgings and the sharpenings and will assist you. It will remember the *results* of the "complex mix of emotions, shapes, colours and textures" that you applied last time, and that will often be what you want. You want to express different things, then, fine, override the AI, the AI will remember what you did, and get better at helping you. I predict that over time digital artists will use the AI tools, the future ones, and it won't even be an issue, and "AI Generated Art", where the AI does most of the work and the 'artist' just lets it, will find its own niche. I'm thinking adverts and cartoons here. You mention book covers and album covers.

This sort of issue has arisen before. When people first produced 'digital art', it was looked down as not real art, but OK for cartoons and kids' first efforts, I can imagine paper artists scoffing "It's just a bunch of coloured squares!" But these days films that use digital effects are praised for the realism of the effects. "Graphic designers" are highly regarded. The digital artist is the latest evolution of the artist, and maybe the 'AI-assisted digital artist' is the next evolution. We cannot tell. It is early days yet.

link Dmytro Nosal'  

From the one side, AIArtists can just type a sentence and have the image. But from another side — will they have sponsors? Will peoples want to pay for another peoples, who just typing sentences and uploading results?

P.S. As I understand, on this image we must not to say "Pepper isn't burning.", as if burning room was normal. We must say "Room is burning.".

link Bene   - Reply

Hahaha, love it. This is fine!

link Haelwenn /элвэн/ :triskell:   - Reply

Haha, ça me démangeais un peu de faire un mauvais troll comme quoi g'mic c'est un peu une AI sur les bords. ^^

Blague à part c'est un très bon article, le coté de virer une méta-donné (le hashtag) d'ailleurs est aussi assez intéressant, ça fonctionne pas que pour des choses comme la vie privée.

link ren   - Reply

I am immensely fascinated by what will be created by the AI ouroboros, as this technology starts to devour its own tail.

But I’ve always been more fascinated by the works that people make, using their own skills and with their own aesthetic choices. It tells a story that AI image generation can never capture, because AI does’t go through the same process.

link eev kuchenka   - Reply

@ziphi absolutely agreed! I like art because I'm interested in the people who make it, not in their technical prowess. And it's cool that people can choose to express themselves through AI art too, but I feel that it doesn't allow the same breadth and originality of expression that full human control does. Sometimes, the crappiest napkin stick figure is more interesting than the shiniest cg painting. :)

link yuvipanda   - Reply

my god this picture is amazing! Do you sell prints?!

link David Revoy Author, - Reply

@yuvipanda Hey, thanks! I used to post things on Redbubble, but they reduced so much the percent for artists that I don't have any incentives anymore to take the time to upload and enter the looooong form to add a new product.

So, just grab the high resolution and print it ( on a USB pendrive; in supermarket there is booth to print big photos, or home printer, or color photocopies does a very good work nowadays).

11 ★

link Adam Dalliance   - Reply

Yeah, as usual the humans are all labeling things wrong.

There is no such thing as AI Art.

There are AI images, but there is no *art* in the images, because art is the product of human emotional connections to the subject, and all there's here is amalgams and machine associations.

Wish they'd just use #aiimage instead of #aiart

link Gumwars   - Reply

@pre I respectfully disagree. AI art is a composite construct of human efforts across the entire race, spanning centuries of work. Art is in the eye of the beholder, and it matters very little what hand guided the brush. What matters more is how do we feel, what have we learned, and where will this new tool of art creation take us? You are focusing too much on the semantics.

link Adam Dalliance   - Reply

@gumwars Yep. Sounds like we do indeed disagree.

Art is what artists do, far from being in the eye of the beholder, the beholders are generally ignorant proles who wouldn't know art if you smacked them around the face with it. 😆

An artist may use whatever they find, including large pixel models, but just typing out a prompt and selecting from the results isn't expression or feeling.

It's not art, it's just iconography.

link Gumwars   - Reply

@pre (p2) Why is AI generated imagery not a valid tool? What makes it less important, less authentic than people literally standing behind a jet engine flinging paint onto a canvas? We are looking at the first steps towards anyone being able to share what they see in their mind to the world. If this isn't a valid form of creation, then art as a medium of visual expression needs to be redefined.

link Adam Dalliance   - Reply

@gumwars If an advertising executive asks David to draw Pepper drinking a lovely can of cola, does that make the advertising executive an artist?

If you want to call these AI images art, then it's far to ask "who is the artist?" in that case.

I can see an argument for the artist being the people who produced all the training data, or maybe the team that built the model.

But definitely not the person who just typed a prompt.

They can maybe be the executive producer or something, but not the artist.

link Adam Dalliance   - Reply

The prompt itself may be art, an expression born of the suffering and emotion of the prompter who wants to tell a story.

But the image isn't art, the image is just a mashup caused by the flow of electrons in response to the art of the prompt, rather than the passion or rage of a human being being human.

link Gumwars   - Reply

@pre You still haven't answered the question underpinning this matter; what is art? Before we can put to bed if AI generated images can be art, or not, you need to define art itself. It isn't fair to say AI art can't exist if you can't define what art is to begin with.

link Adam Dalliance   - Reply

@gumwars Yeah I did, I said it's what artists do.

Lying in bed all day can be art if an artist does it.

Good art is born from human passion and experience, it's an expression of their frustration and pain and joy and curiosity.

Which are all things that the machine hasn't got, and which preclude it from being an artist, and so preclude it from making art.

That's these machines that are just mashing up pixel technique anyway. It may be possible one day to build a machine which can experience joy and curiosity and longing and which is about to express that.

But these machine's aren't it.

The art of AI images is in the prompt writing, and I would rather read the prompt.

link Adam Dalliance   - Reply


And an artist may use these machines in their work. If they have a story to tell, and illustrate the story with photographs or a collection of pebbles or a film of a plastic bag floating, or the pixels that fall out of a pixel grinder that's all cool and legit.

But the art is in the story there, the human communication, mind to mind, not the pebbles or pixels.

link Gumwars   - Reply

@pre Art is what artists do? Dude. Let's reductio ad absurdum this a minute. You're telling me if an "artist" takes a dump on a miniature Eiffel Tower and calls it "Brown Spike" that it's now "art"? Then by that same reasoning, if an "artist" prompts MidJourney "Photo of the Eiffel Tower, covered in poop, 4k, extremely detailed, by Greg Rutowski" it now becomes art? GTFO. That is the most lazy and absurd definition I've heard, of nearly anything.

link Adam Dalliance   - Reply

@gumwars That is indeed what I'm saying, but you missed the subtle difference.

When an artist prompts midjourney about his poop covered french monument, it is the *prompt* which is the art, and midjourney's output is derivative processing from that art.

Anyone can be an artist! They just need to have the intent to produce art and then make something up to express themselves.

Whatever they make will be art!

What you type into stable diffusion is art because an artist made it!

What comes out of stable diffusion isn't art because a person didn't make it, it was formed from matrix multiplication processing upon the actual real art of many painters and one prompt writer.

link Gumwars   - Reply

@pre But that image would not have existed if not for the human hand that; created the model used to generate the image and the words given to the model to create the image itself. Every instance AI is used to create an image is something entirely brand new, never seen, a combination as infinite as there are stars in the sky. Saying that it isn't art ignores the fact that it can be emotional, inspirational, and every bit as compelling as work that is crafted using older ways.

link Adam Dalliance   - Reply

@gumwars A sunset can also be as emotional and inspirational and compelling as actual art made by actual artists.

But a sunset is not art.

Unless god is an artist maybe.

link Gumwars   - Reply

@pre True, but a sunset is not guided by human hands. If someone with no formal training in art can sit down in front of their laptop and create an image that moves me, that's art in my book. Looking back to when abstract expressionism and neo-abstract art became a thing, the question was immediately raised regarding the qualification of the people essentially slapping paint on things and charging 5 figure sums. The response is, if it moves you, it doesn't matter. That is art.

link Adam Dalliance   - Reply

@gumwars Hehe. Maybe where you live, but I live in London and that horizon is made of buildings which are art and the brilliant red sky is made of pollution which isn't.

So the sunsets are guided by human hands.

But it's *intent*, that makes the difference between the dome of StPauls and the smog red shine of the sky.

To the extent that the image is indeed intended by the prompt engineer it is art, but every single bit, every single 1 or 0 of that input is in the prompt.

If a Stable Diffusion rendering on the phrase "some art" is art, then who is the artist?

link Gumwars   - Reply

@pre Have you played with stable diffusion much? Prompting, control net, DAAM, ROOP, negative prompting, model selection, Chainner, figuring out how to upscale without distortion, outpainting, inpainting, prompting the out/inpainting, selection of latent space, depth masks, ...I'd argue the toolbox is pretty complex and growing by the day. Yes, there is plenty of AI imagery that is garbage. I'd also argue that some is absolutely art, taking skill to create.

link Adam Dalliance   - Reply


I made this thing where I told it to cut out my face from my avatar and render new movie-poster backgrounds all night long then stitched it into a video.

There are a lot of obvious copyright infringements in it, some actual trade-mark violations, and barely two hours work from me.

I feel like it told the story I wanted told, of how cookie-cutter and templated most of modern movies are, and how cliche and formulaic and repetitive the output of the machine tends to be as a result.

I wouldn't expect anyone to watch it to the end. I didn't.

I wouldn't call any of those images art, but maybe the whole project as an entirety is.

I am certain there is a skill to making the images into what you want them to be, especially if you feed them the right blobs as init-sketches.

The more you know what you want before you start, the more it's art. The more you just feed gags into a machine to see what happens, the less so.

The beauty of those things is in the eye of the beholder, but the art is in the intent of the artist.

link theo⏚ ✅   - Reply

Nice thought-provoking post.

Although, I'm not convinced that it's an issue for an AI to use data generated by other AI.

Just use near-slaves in 3rd world countries to to teach your system to generate something a human would expect.

link Goofy 📖 🍝 :unverified:   - Reply

C'est bien intéressant, en particulier sur ton mode de création qui semble passer directement de l'imagination à la réalisation dessinée sans passer par la verbalisation textuelle.

link Adam   - Reply

Great post! You're hard work, human touch and meaning is what provides your art value. You will prevail.

link Simeon Nedkov   - Reply

Good call!

link boingo-00 🦊 :fluffytail:   - Reply

I tried Stable Diffusion and other neural networks to generate some drafts. Of course it is far from perfect, but it is good instrument to look if one or another scene does ever make sense
I want to learn how to draw, but until i can, i have a tool for playing with my thoughts

link David Revoy Author, - Reply

@me yes, I collected also feedback about how AI image generation might bring more people to digital painting: to learn at first how to fix imperfection on their pictures, then out of frustration to not get the result they want learning how to make them the manual way.
That's a new and interesting entry point to digital art.

link Olivier Saraja ☕️📚🦖🛸👻   - Reply

I feel this as the right move. Not the easier, but the right one.

It is not to the humans to provide proof of their genuine work, but to the AI generated one to be tagged as.

You have all my support, old friend!

link David Revoy Author, - Reply

@oliviersaraja Thank you Olivier!

2 ★

link Olm-e   - Reply

@oliviersaraja I agree in the argument
The label I made was a sort of satire : it's totally insane reality is now turned over upside down and people have to justify their art is human made or to restrict the use in predatory database ... I think laws and practice are evolving and hopefully dust will settle and we'll get back to more reasonable conversations
but still, artists needs to organise more to defends collectively their rights, visual artists are strangely not unionised

link Pixel_Lime   - Reply

This is how I feel in the summer because of the hot weather.

link David Revoy Author, - Reply

@Pixel_Lime 🤣

I hope you'll have many fresh days this season!

link chrismckee   - Reply

love this version of the 'this is fine'. My daughter loves your comics

link David Revoy Author, - Reply

@chrismckee 😍 thank you!

link Ivan Sagalaev :flag_wbw:   - Reply

completely agree with your choice. Actually creating things is what's normal, it's re-purposing of existing work that should require special attribution.

Thanks for doing your part of keeping the Web human and interesting. I enjoy your work!

link LumiWorx   - Reply

... speaking as a creator for 60+ years, most of us create because we must. It isn't rooted in money or fame, or from a request. It is an imperative that sustains our soul and our very human existence.

What we create - "art" included, in every sense of the word - is born from human expression and personal experience. A Spirograph is not an artist because it mimics a human with a pen. A CPU has no form of expression, nor code, nor a pile of gears. Art is not "the process"!

link David Revoy Author, - Reply

@lumiworx @gumwars

> It is an imperative that sustains our soul and our very human existence.

Oh yes, and it took me time to realize it was a necessity to me for reaching a psychological balance.

3 ★

link Gumwars   - Reply

@lumiworx Absolutely. I write because there are stories in my head that need to be told (even if it's only to the walls of my room). I draw because those stories need to be visualized. I paint and model because I need to see those characters in this world, not just on a paper or canvas. I use AI to iterate those ideas. I start with work that I've created and explore it in ways that would take me months to do otherwise.

link LumiWorx   - Reply

@gumwars what you describe seems to be a collaboration if the handoff was to another person. I won't speak for anyone else or some abstract point of view...

If I were to do the same thing, my creation would no longer be 'mine'... it would turn into someone elses vision and interpretation, and out of my control to match what I alone see. I just couldn't put my name on it.

I have been refining 'her' for 15 years, and I'm still not done. Time invested is irrelevant.

link Gumwars   - Reply

@lumiworx Do you attribute the software engineers that created Blender, Maya, or Zbrush in the creation of your work? Or are those simply tools you used to actualize your vision? Granted, a truly hardcore 3d artist could create an image using plotted points and a .CSV file, but that seems a pretty harsh hurdle to get over to "own" a particular work.

link LumiWorx   - Reply

@gumwars The 3D model would be mine, as would the render created to present it.

Even though I'm not required or asked to... Yes, I do attribute the software when I present a render, commercial or FOSS.

However, if I were to publicly show a painting or a drawing or a film print, no, I don't credit the paint's or camera's maker, nor do I feel their branded technical expertise in their products is relevant. They are simple non-thinking tools.

link LumiWorx   - Reply

@gumwars I'm not suggesting anything is wrong with your approach. Visualizing another step past where something was can be a valuable tool, but for me, that wouldn't be my final output.

link Gumwars   - Reply

@lumiworx Nor is it mine. I'll take a character sketch, plug it into SD, and can create 100 iterations in about 10-15 minutes. I can explore different palettes, textures, backgrounds, or just riff on ideas. I can be vague or specific. Those go back under my pen (or tablet in this case) and I work on them further.

link LumiWorx   - Reply


All I could offer is that #ai is not in any workflow I personally would use, and I can say with certainty that it will not be a part of it in the future. I say that with no reservation and with no judgment.

#Pollock could have been accused of being an 'accidental artist' because he chose not to use a brush and spill paint on a canvas. His work wasn't accidental, but he embraced an errant droplet or dribble of paint as long as it kept his vision intact.

link LWFlouisa   - Reply

Your art looks awesome.

Glad other people are getting tired of the digital == aiart.

link David Revoy Author, - Reply

@LWFlouisa Thank you

link LWFlouisa   - Reply

I wonder if eventually if AI art consumes to consume its own byproduct, if that will eventually collapase that industry.

I make AI myself, but I very specifically such that it doesnt do web scraping and other unethical things.

So hearing FOSS is doing it gives me reason to hope.

link Stemy   - Reply

Avec les NFT et l'art IA, il ne fait pas bon d'avoir de l'ambition artistique ces derniers temps. Je vois se profiler un monde où tous les artistes humains auront disparu, où le seul semblant d'art produit ne sera que le résultat sans âme de quelques lignes de code, ce que je qualifierai de dystopie. ça me console un peu d'apprendre que l'art AI se prenne toute cette haine. Bon, ça fait très OK boomer, mais tant pis.

link David Revoy Author, - Reply

@stemy Oui, je comprends en tout cas si ce contexte décourage certains jeunes artistes, ou les pousse vers l'aquarelle et les techniques traditionnelles. C'est l'informatique tout entier qui risque de se priver de talents.

link Evv1L (Эвил)   - Reply

literally said: :blobcatthisisfinegoogly:

Good art 👍

link Agonio   - Reply

ty for the blog post, an interesting read. I definitely agree with the problem of consent for data usage, it's an abuse of the privilege of companies with high server-power.

Just a note. You start by saying that you see people mistaking human art for AI, not able to appreciate the effort when it is there.
I believe this misses the issue. I don't think art is something you make to show off effort and skill... We need to move past that. This should resonate with the end of your post.

link David Revoy Author, - Reply

> not able to appreciate the effort when it is there.

Yes, you are right. It shouldn't be related to effort or skill. But I couldn't refrain to write it, probably the frustration of spending all my hours and years of experience to get my paintings and being compared with someone texting something.

Thanks for the comment.

link ArtfulGenie   - Reply

I am a fellow creator myself and have given this much thought. I believe that very soon companies and individuals looking to profit off the use of AI-generated content will be federally regulated or even universally regulated and will be enforced to say this advertisement for example is 50% AI generated. Some companies are already doing this. That being said, I believe that you are ahead of the game by using hashtags such as #realartist #realart #humancontent #realartisty #humanlyinspired, and soon #humanartsfoundation That being said, I feel as if starting a movement that aims to become an open source platform that serves the interests of the artistic community and those that want to support it and be able to find art that is purely created by humanity. Now, a very high-tech algorithm would have to be in place to detect AI-generated content and it would be non-profit. I believe it will take the artistic community to form this movement before it is implemented by big tech, rather than big tech being ethically involved. The “human” art movement will continue to grow into an even bigger industry and I want to ensure that we as artists have as much control as we can going forward. The human art foundation so to speak, if the name is implemented and agreed upon by the community, will include all art forms - poetry, writing, video, music, songwriting, and yes even digital art and graphic design. I would love to collaborate with you and have a full conversation please let me know how we can possibly get in touch and start this movement !

(Original message has been truncated: read the complete original message here.)

link David Revoy Author, - Reply

@ArtfulGenie Hey, thanks for your nice words. Welcome to the Fediverse, I checked your profile/page to try to know more about you but I saw you just joined and you still have to fill up your profile. It will be easier for me to meet you first in this way. Speaking to collaborate or engage in a full conversation is a bit early 😉

link Owl!🦉   - Reply

> But lately I started to enjoy this behaviour in a new way, because this faking could be what wastes all the data sets and training models

AI artists from civitai have already made a million of loras, trained on the midjourney outputs. For now it's more like compiler bootstrapping rather than "eating ourselves".
To be honest, I think the main problem with using generated images in training is that someone would say that even though the datasets contain no copyrighted hand-drawn images, it's still "unethical" , because they are generated using old models and continue this "ethics" shitpost.
In order for this to make sense, we first need to create some kind of model to which no one will have any complaints, and then start bootstrapping.

> I will continue to publish all under a permissive Creative Commons licence and with the source files, because that's how I like my art: free and libre.

Just at the moment when computer geeks came up with idea how they can use your images to create somewhat free and open source it turned out that it's forbidden. This is funny.
I understand why you may not like midjourney, firefly, dalle2 and other non-free stuff (i don't like them too), but SD is completely free, open source and can be used or modified by anyone...

> With AI, AIArtists simply type in keywords for a subject.

No, we don't, stop it. With no control and editing at all ai draws random broken shit.

(Original message has been truncated: read the complete original message here.)

link David Revoy Author, - Reply

> SD is completely free, open source
Yes, but not Free/Libre as GPL (src )

> it turned out that it's forbidden.
it's not, but if you do derivation of my art, you have to keep attribution. I should get credited in this regards legally in all AI images, because my art is part of the dataset. (but I don't want, that would be a nightmare). Also, got already ethical request to train on my art and accepted 😉

link Owl!🦉   - Reply

> Yes, but not Free/Libre as GPL

It's similar to more permissive licenses, i don't think this is bad.

> I should get credited in this regards legally in all AI images

This is technically possible, pic related.
However, result image isn't even bareilly similar to your drawings, also model contains information about thousands of other artist's drawings, which maybe or maybe not used to produce this shitty example test image...

🖼️ original 

link Owl!🦉   - Reply

Among other things, your problem is that you reduce the whole process of using AI-art models to "entered three words - waited 8 seconds - got a picture", but this is not always the case. To be honest, I rarely use ai this way.

What do you think about this images, are they ai/unethical or not 🤔

link David Revoy Author, - Reply

@lonelyowl Please feel free to publish articles, videos and tutorials about your process if you think your audience needs to be informed about your craft involving AI tools. I might reconsider my words about the simplicity of entering a prompt and hitting enter.

About the ethic, I have no idea what you are asking... what you used precisely? based on what dataset? and what you want to make with it? Also: not my job to comment on that, if you have an ethic, it's your job to communicate them.

link Owl!🦉   - Reply

I hardly have the patience for full-fledged tutorials or vids, but I think that I could at least do step-by-step processes for creating these pictures.
To be honest, I'm not particularly interested in organizing some ai art-education, I just want to stop being considered a subhuman.

> About the ethic, I have no idea what you are asking... what you used precisely?

The line to the pictures in this post was drawn by hand in procreate and colored by using a controlnet/canny, the model used is camelliamix_line v2 ->

link hairylarry   - Reply

I think they're very nice.

link Sybren Stüvel 🔶   - Reply


link David Revoy Author, - Reply


link Resuna   - Reply

If we ever develop AI this will no longer be an issue. Calling generative neural networks "AI" makes my brain itch.

link hairylarry   - Reply

I love the insight into your thought process in this article.

link Badri   - Reply

I missed the discussion, but I love the idea of "AI eating itself"! I wrote an article speculating that this could happen, a few months ago 🐍

It was exciting to see the Platformer piece you linked to talking about the same ideas! I didn't realise it's already happening though, but it's true; I don't trust Google (or even DuckDuckGo) search results the way I used to, and I'm having to scroll a lot to find genuine content 🙁

Your art is very recognisable though and those who know you know it's certainly not AI. Also, I noticed you put the #krita hashtag here—for those who know what that is I suppose that's the most self-explanatory 😉

link ジョアン   - Reply

I will continue supporting you and your art, but, at this point, I don't think there is much anyone can do to stop the AI art. I'm glad you're not stopping

link VulcanSphere   - Reply

@davidrevoy Thank you for your decision, deevad

As a Pepper & Carrot fan and AI art/synthography fan, Vulcan thinks we need to shift our paradigm and treat generative art as helpful tool for artists. A generation is not perfect and retouches (outpainting, inpainting, manual post-processing with graphics editor) are still needed to yield a decent artwork.

And generative art is still generated by humans, there is no sentient robot yet.

link der.hans   - Reply

Yes, we will continue to support you!

Thanks for sharing your art with us :)

link David Revoy Author, - Reply


link West Lawns   - Reply


Post a reply

The comments are synchronised every 12h with the replies to this post on Mastodon:

How to use this? (click here to unfold)
Open a new Mastodon account on the server of your choice. Then, Copy/Paste the adress above in your Mastodon 'Search' field. The post will appear and you'll be able to fully interact with it. You'll have full control of your posts: edit, remove, etc. After that, your message will appear here.

Just please note that it may take up to 12 hours for your changes to be reflected here.