I’d like to add that even without AI, art is already over-saturated on social media platforms in the sense that you can expect most people to look at the piece you poured dozens of hours into, for a second or two before they scroll to the next thing. Maybe leaving a thumbs-up if they liked what they glimpsed. And because algorithms rewards users more from drama than great art (clickbaits, reactionary posts that people engage with, very frequent posting etc), it’s harder and harder for new artists to get their works seen.
With AI generators, art is more and more seen as fast-food content rather than an expression of human creative skill and imagination. And that worries me personally, as I don’t think it’s going to be a great change for humanity to suck humanity out of arts.
I believe AI is a tool to be used
I don’t think AI is art theft, at least not inherently.
The theft is mostly on humans, lying and passing off the work as their own.
My biggest concern about it is mainly how fast the tech is developing and how slow humanity is in responds to rolling out laws about it.
AI voice, picture manipulation, video manipulation, writing etc. lots of room for malices acters to abuse it and we need rules for that.
However using one of the Art AIs out on the market and ask it to create a mismatch of data into something vaguely in line with your desired outcome shouldn’t be immoral I think.
The immoral act is in the intent to either sell it or pass it off as your own creation.
Least that is what I think
There is plenty of arguments to be used against and for my reasons, and I could write a longer wall of text going into depth for why I think this way, but I rather not get into a heated argument on such a divided topic.
“Arguing on the internet is a game where the best move is not to play”
I seems to me X/twitter is a poor platform for Art anyway, with its rampant toxicity and drama focus
I doubt tictoc and instergram is better for digital art specifically.
and Facebook is… I don’t know what to think of that.
There is tumblr but its user base is small?
Are social media platforms (the big ones) actually accommodating to digital and pen drawing artists at all? and if not would it not be better to migrate to another site?
Artstation perhaps?
Absolutely abhorrent. It’s theft of other artists using shoddy convenience tools because of either laziness or reluctance to want to pay for someone’s work.
I understand that not everyone can afford £70 + commissions for art of their character, especially new characters that they aren’t even sure if they’ll keep RPing. But there are a plethora of FREE art resources out there that act as wonderful placeholders without simultaneously stealing the work of others, such as Heroforge, Picrew, etc.
Additionally dozens of artists hold art raffles, and I personally have won a few art pieces just by being an active member in online art communities. And in doing so, you are supporting artists actively by boosting their accounts and increasing their reach.
There is no excuse for AI art anywhere. Not ever forever. Not only is it inherently lazy and stingy (and also makes no sense, for a yearly Midjourney subscription you could easily afford a piece of custom art from an artist) but AI generative models are inherently plagiaristic in nature with how they scour, sample, imitate and mass produce ‘art’ using the work of other people.
Nothing original comes out of AI generative models, they are not human, they have no creativity. They are algorithmic and stands against everything that actual artistry is.
sorry for big rant, I have very strong feelings on this subject lol- but yeah, entirely against it, and if I see AI character art in someone’s about then I am going to stay a million miles away if I can manage it.
If an artist wished to train a model entirely on their own works, then no, that wouldn’t be theft.
But every AI model out there uses a dataset that’s built from scraping twitter, deviantart, artstation, etc. for other artist’s work.
So yeah, they’re built on theft.
One need look no further on that than stuff like getty images suing stable diffusion https://www.theverge.com/2023/2/6/23587393/ai-art-copyright-lawsuit-getty-images-stable-diffusion
Kelly McKernan - the artist who Acrona linked earlier - is also joining with other artists for extremely understandable reasons. https://www.theverge.com/2023/1/16/23557098/generative-ai-art-copyright-legal-lawsuit-stable-diffusion-midjourney-deviantart
Yup – when a tool requires theft in order to function, it doesn’t matter how honest the user is about it.
Amusingly, I’ve seen some stuff about image generators scraping other generated images since they’ve become so widespread, making the final product even worse.
It feels antithetical to engage in having a machine generate something for you in a hobby entirely dedicated to enjoying the creative process of writing and drawing.
I’d rather just see a screenshot of someone’s character than anything AI generated. In the WoW RP “scene” we are lucky to have a lot of talented creative people who breathe life into our roleplay through their art, so aside from the arguments against AI generated art that have already been made here, using it seems to me like a kick in the teeth against the people who are often our friends, guildies and fellow roleplayers.
I don’t want to get into a long drawn argument here
but humans do that all the time
are we giving them a pass because they take longer to copy the work?
where are we drawing the line?
I agree in so far AI should be limited in how much it can access, but I do not persoanlly see how it fundamentally work as any different from how humans do it,
again I think the maliciousness lies in the human not the AI usage itself.
You see this in academia as well. The collective attention span of people in the below-30 bracket is plummeting overall, which directly impacts the quality of their learning and research capabilities.
elaborate. Do you mean human’s stealing other people’s work? Tracing to pretend that you did something yourself?
Cause yes, that is bad, should not be encouraged, and is something that all respectable artists everywhere are against.
But if you mean art referencing or tracing for practice then politely be quiet because they are not comparable even remotely.
Machine-generated imagery, commonly mislabelled as “AI art”, is theft and I will burn bridges with anyone who partakes.
Fortunately we banned it completely from the Vigil recently.
We do not recognize machine-generated imagery as art and will refrain from referring to it as such, nor is anything about machine-learning-models actually intelligent. This is relevant because machine-generated imagery requires minimal creative human input, no training and is unable to reflect the human spirit that makes art unique. Legally, an image-generator cannot hold copyright over its own generations and the “prompter” (who instructs the model how to generate images via text-based prompts) takes the role of a commissioner. Since the machine has no copyright of its own generations, it cannot transfer the copyright to the commissioner.
This is funny in so far as that lately these machine-prompters, who previously laughed at artists who complained about their art being used, complain that the images they commissioned (generated) as well as their prompts can be used with impunity by other promptoids.
The datasets used to train all major image-generation models, Laion 5b, consists of over 5.85 billion images that amount to over 60 Terabytes of images and metadata that were scraped indiscriminately from the web without asking for the authors’ consent.
A large portion of this data is copyrighted art, photography and illustrations, as well as sensitive private information such as medical records. This constitutes illegal copyright- and personal data infringement.
Machine-image-generation is directly marketed and adopted to deprive human artists of work.
As of yet, machine-image generation is largely unregulated by international copyright and thus provides no legal precedent for plagiarized artists to litigate on.
Using machine-generated images as inspiration to base your own art on from scratch is one thing, but taking a generated image and displaying it as your own creation is an affront to actual artists all over the world.
I would like to offer a small glimmer of hope among all the doom and gloom.
AI is, in fact, not profitable. Open AI (behind ChatGPT) and Stability AI (behind Stable Diffusion) are both currently going through serious management issues due to investor unhappiness on returns on their investment. The latter is reportedly seeking someone else acquire them. https://www.pymnts.com/acquisitions/2023/report-stability-ai-positioning-itself-for-acquisition/
Techbros invested into this the same way they did in NFTs, only to get hit by the inevitable realisation that actually no, it’s not quite the golden goose.
We do not give a pass to people who plagiarise, actually. In an academic setting it gets your hand slapped hard, in an artistic setting it can lead to being shunned.
And if it plagiarises in a way which infringes on someone’s IP rights, it can become criminal.
We don’t give humans a pass on plagiarism, either socially or potentially legally. Why are you giving a machine a pass to do it en masse?
Ah so there is a difference?
I was not aware.
to me it seems to be the same thing, only the intended outcome is different, one being innocent pursuit of improvement and the other more malicious.
Computer-generated imagery is, in fact, very immoral from a Kantian point of view. It breaks with the principle of humanity, which states that you should treat other people as a goal in themselves, not as a tool. Basically; you need consent.
Computer-generated imagery is immoral because it does not ask for consent. When you use a CG tool, it uses the work of an untold number of artists, without consent. Therefore, you are acting immorally towards and untold number of people.
If there was an image generator trained entirely on a closed dataset, where all the artists consented, that would be moral. As far as I’m aware, that does not exist yet.
I will also say that, in my mind, computer-generated imagery is aesthetically wrong. It just looks ugly. Every image is the same vague, blurry semi-coherent mess, with no intent or creative spark motivating anything. Avoid at all costs.
I can see AI art becoming a tool, an aid if you will, to proper artists and similar creators. No point nor reason to simply give up and pack it in tbh.
They used to be much better only a few years ago. Tumblr doesn’t get as much traction as it used to. Instagram? This video summarizes the issue for new artists pretty well:
Artstation? Artists started protesting when AI images started flooding that site too:
AI stuff started also flooding Artstation Marketplace where artists could sell their products. It’s also more of a professional portfolio site, while DeviantArt used to be where hobbyists and art appreciators mingled plenty, but it too has been flooded by AI images and I rarely use it anymore.
Whenever I try google art references for inspiration or learning, half of the search results are AI generated images. At least I can still (for now) find actually human made art through Pinterest.
So as a whole, AI imagery is a nuisance. I’m not opposed to new tools that help creatives do their job, and professional artists today have to keep learning new tech all the time, but this is all about companies data laundering artists to profit from a product that competes with artists in their own field.
