3D Models
3D Print Models
3D Scans
Animation & MoCap
Textures
Materials
Skin Textures
2D Game Assets
Sound Effects
Brushes & Tools
Reference Photos
Stock Images
HDR Images
3ds Max
Maya
Blender
Unreal Engine
Unity
ZBrush
Cinema 4D
LightWave
Marmoset
SketchUp
FBX
DAE
DXF
GLBWhat is your proposed solution to AI??
430Thread Activity
AnabranWed, Apr 16
AnabranWed, Apr 16
AnabranWed, Apr 16
AnabranTue, Apr 15
AnabranTue, Apr 15
guy91600Tue, Apr 15
Serious question.
As there seems to nearly universal agreement
the generative AI technology will not be uninvented.
I would like to hear realistic proposals on how you would address the challenge of its existence going forward as an individual
when I say realistic I mean what could be done
within the confines of current laws in western society.
1) Emotional appeals to change heart and minds of non artists laypeople who do not seem to care at all how art is made only that it is consumable?
2) More lawsuits against the AI companies to get money from them somehow.
3) Total boycott of any creative software company that uses AI in any form?
(such companies include)
Daz inc ,Adobe, Canva/Affinity, Black magic designs Davinci Resolve, Vegas pro, Cascaduer, Reallusion, Topaz, capcut
and many many more.
Renderhub Recently prohibited images "substantially created with AI"
without providing a hard definition of the term "substantial."
Whenever this subject is discussed
I see the same slogans about supporting human artists
how are supporting them beyond repeating such slogans?
As there seems to nearly universal agreement
the generative AI technology will not be uninvented.
I would like to hear realistic proposals on how you would address the challenge of its existence going forward as an individual
when I say realistic I mean what could be done
within the confines of current laws in western society.
1) Emotional appeals to change heart and minds of non artists laypeople who do not seem to care at all how art is made only that it is consumable?
2) More lawsuits against the AI companies to get money from them somehow.
3) Total boycott of any creative software company that uses AI in any form?
(such companies include)
Daz inc ,Adobe, Canva/Affinity, Black magic designs Davinci Resolve, Vegas pro, Cascaduer, Reallusion, Topaz, capcut
and many many more.
Renderhub Recently prohibited images "substantially created with AI"
without providing a hard definition of the term "substantial."
Whenever this subject is discussed
I see the same slogans about supporting human artists
how are supporting them beyond repeating such slogans?
! REPORT
I have no idea, what realistic might be in the end. I don't know enough about the technology behind AI.
Is there a way to lock AI products to its own specific file format, in order to be identified as AI?
Can ther be a sort of digital (invisible) watermark, that can be read out?
Part of the danger of AI is, that it might not be able able to be recognized as AI.
Is there a way to lock AI products to its own specific file format, in order to be identified as AI?
Can ther be a sort of digital (invisible) watermark, that can be read out?
Part of the danger of AI is, that it might not be able able to be recognized as AI.
REPLY
! REPORT
It should be confined to a tool-based solution that enhances humanity, not competes with it, especially since the only way it can do so is by taking from humanity in the first place.
REPLY
! REPORT
I still believe the most realistic approach to controlling these products is by controlling what is allowed in the libraries they are trained on. In the US, NIL (name image likeness) laws are already in the books thanks to lawsuits demanding recompense for NIL use in video games.. and let's not forget actual copyright laws.
Lawsuits over NIL and/or copyright infringement result in a digital discovery phase that opens up the training libraries for inspection. If sufficient legal pressure is applied, a natural side effect is to catalog what is in the library for quicker resolution. And this is where I believe the control starts.. cataloging, audit, and risk assessment against training libraries, risk reduction actions should result in buying art pieces instead of just scooping them up and the ability to "flag" (and remove) samples in a library as stolen content.
Lawsuits over NIL and/or copyright infringement result in a digital discovery phase that opens up the training libraries for inspection. If sufficient legal pressure is applied, a natural side effect is to catalog what is in the library for quicker resolution. And this is where I believe the control starts.. cataloging, audit, and risk assessment against training libraries, risk reduction actions should result in buying art pieces instead of just scooping them up and the ability to "flag" (and remove) samples in a library as stolen content.
REPLY
! REPORT
guy91600
Karma: 12,988
Sun, Apr 13"The ability to "flag" (and remove) samples in a library as stolen content."
This is what should be done, but it's a dream when you consider how difficult it is for an artist to get hosting platforms (Google Drive, Mega Drive, for example) to remove products uploaded by pirates.
This is what should be done, but it's a dream when you consider how difficult it is for an artist to get hosting platforms (Google Drive, Mega Drive, for example) to remove products uploaded by pirates.
I think, it's a pointless discussion, not a single human being here in this forum is able to do something against it, change any laws nor create any laws.
What I would like to see is, that it's clearly stated, how something, be it a render or what ever, is made.
Best with a huge watermark on any AI products, that can't be removed, so that the work from humans is safe and will be paid better, than any fast made prompt generated whatsoever.
Just for a picture, who wanna have a picture at the wall with a big watermark all over it? Not me and I would say, no one else either.
Or music with some kind of noise in the middle of it or a movie and so on .... I think, you got the picture.
What I would like to see is, that it's clearly stated, how something, be it a render or what ever, is made.
Best with a huge watermark on any AI products, that can't be removed, so that the work from humans is safe and will be paid better, than any fast made prompt generated whatsoever.
Just for a picture, who wanna have a picture at the wall with a big watermark all over it? Not me and I would say, no one else either.
Or music with some kind of noise in the middle of it or a movie and so on .... I think, you got the picture.
REPLY
! REPORT
Luxe Muse
Karma: 5,150
Mon, Apr 14I hear you. I kept my response short and slogan-like because I know we're in a bit of an echo chamber here.
I really like what you shared, and it got me thinking. Maybe we can't do much about the source of training data (images, chats, books). But the output pipeline, the tools used to run LLMs, is a smaller target. They could weave blockchain technology in deeper, recording and digitally watermarking every AI creation. That immutable blockchain could act like a footprint, letting us verify if anything matches an entry in the chain.
I really like what you shared, and it got me thinking. Maybe we can't do much about the source of training data (images, chats, books). But the output pipeline, the tools used to run LLMs, is a smaller target. They could weave blockchain technology in deeper, recording and digitally watermarking every AI creation. That immutable blockchain could act like a footprint, letting us verify if anything matches an entry in the chain.
Beautiful Misfits
Karma: 16,041
Mon, Apr 14Thank you 
I wrote my real answer a bit further down, cause this here was just something I would like to have, since I know, AI will stay.
But honestly, it doesn't bother me at all, cause I see behind a pointless discussion, I see a bit further into a future, I don't like.
If copyrights will be our only concern about AI, oh heaven, I would be happy.
I wrote my real answer a bit further down, cause this here was just something I would like to have, since I know, AI will stay.
But honestly, it doesn't bother me at all, cause I see behind a pointless discussion, I see a bit further into a future, I don't like.
If copyrights will be our only concern about AI, oh heaven, I would be happy.
Wait and see if it's something that's actually here to stay or just some passing trend that has its moment.
I think it could remain - or develop into - something that's used to fill some creative gaps, but the whole "Make something awesome for me!" bit will probably fade out into some sort of niche interest.
I think it could remain - or develop into - something that's used to fill some creative gaps, but the whole "Make something awesome for me!" bit will probably fade out into some sort of niche interest.
REPLY
! REPORT
Quote: What I would like to see is, that it's clearly stated, how something, be it a render or what ever, is made.
Best with a huge watermark on any AI products, that can't be removed,
The problem with technology based counter measures such as the watermarks you suggest, is that unless you actually change your laws to make using the tech completely Illegal without prior government approval, you likely would not have any legal basis for enforcing a watermark obscuring someones legally created AI film ,illustration etc.
At best you could hope for is a requirement to disclose the use of AI which google already requires when uploading a video.
But I remain unconvinced that the general pubic who pays to see/stream movies cares enough to be discouraged by such disclaimers from watching visual Entertainment anymore than they are by the current rating system warning of violence, profanity in R rated films etc.
Best with a huge watermark on any AI products, that can't be removed,
The problem with technology based counter measures such as the watermarks you suggest, is that unless you actually change your laws to make using the tech completely Illegal without prior government approval, you likely would not have any legal basis for enforcing a watermark obscuring someones legally created AI film ,illustration etc.
At best you could hope for is a requirement to disclose the use of AI which google already requires when uploading a video.
But I remain unconvinced that the general pubic who pays to see/stream movies cares enough to be discouraged by such disclaimers from watching visual Entertainment anymore than they are by the current rating system warning of violence, profanity in R rated films etc.
REPLY
! REPORT
Beautiful Misfits
Karma: 16,041
Sun, Apr 13Well, as I said before, it's pointless and not for the obvious reasons.
Those discussion only help to keep the people entertained, so they won't see the real problem coming.
And I'm telling you, copyright or not, any new laws or not, all this doesn't matter at all.
We are facing much bigger problems, the elephant is standing in the room and no one can or will see it, instead discussion about AI and real art *lol*
This here will be our smallest concern. And yes, you're right with that point, the general public don't care and don't mind.
Popcorn and movies *hurray* (bread and games) and no reason to think just for a second further.
The Sorcerer's Apprentice
Oh, here comes the master! Lord, the need is great! I cannot get rid of the spirits I called.
Only that there will be no master to stop that, what's coming.
Enjoy the show, be part of it, if you want, every training is welcomed by those, that will control.
This discussion is so freaking pointless!
Those discussion only help to keep the people entertained, so they won't see the real problem coming.
And I'm telling you, copyright or not, any new laws or not, all this doesn't matter at all.
We are facing much bigger problems, the elephant is standing in the room and no one can or will see it, instead discussion about AI and real art *lol*
This here will be our smallest concern. And yes, you're right with that point, the general public don't care and don't mind.
Popcorn and movies *hurray* (bread and games) and no reason to think just for a second further.
The Sorcerer's Apprentice
Oh, here comes the master! Lord, the need is great! I cannot get rid of the spirits I called.
Only that there will be no master to stop that, what's coming.
Enjoy the show, be part of it, if you want, every training is welcomed by those, that will control.
This discussion is so freaking pointless!
guy91600
Karma: 12,988
Mon, Apr 14@Beautiful Misfits
Maybe a small number of readers of this discussion will open their eyes after reading your opinion, at least I hope so
Maybe a small number of readers of this discussion will open their eyes after reading your opinion, at least I hope so
Beautiful Misfits
Karma: 16,041
Mon, Apr 14Thank you guy, I'm happy that I reached one ... You
I'm afraid, no one else will, hiding the head in the sand is so much easier (ostrich-politics).
I'm afraid, no one else will, hiding the head in the sand is so much easier (ostrich-politics).
Quote: Wait and see if it's something that's actually here to stay or just some passing trend that has its moment.
I think it could remain - or develop into - something that's used to fill some creative gaps, but the whole "Make something awesome for me!" bit will probably fade out into some sort of niche interest.
The underlying tech is too utilitarian to just fade away IMHO.
But the many secondary industries exploiting the AI hype train ,that you see all over youtube,
(How to make $500 a day with chatGPT!!) ,will burnout when it becomes apparent they offer no actual beyond AI hype.
I think it could remain - or develop into - something that's used to fill some creative gaps, but the whole "Make something awesome for me!" bit will probably fade out into some sort of niche interest.
The underlying tech is too utilitarian to just fade away IMHO.
But the many secondary industries exploiting the AI hype train ,that you see all over youtube,
(How to make $500 a day with chatGPT!!) ,will burnout when it becomes apparent they offer no actual beyond AI hype.
REPLY
! REPORT
Quote: "The ability to "flag" (and remove) samples in a library as stolen content."
There are no samples to remove from any libraries
CHKPT training data bases does not store any original images
just some code that represents its "knowledge "of how to create images based on LLM text prompts
:text=Diffusion%20models%20work%20by%20adding,order%20to%20produce%20new%20outputs." target="_blank"~https://arxiv.org/html/2312.14977v1#:~:text=Diffusion%20models%20work%20by%20adding,order%20to%20produce%20new%20outputs.
Quote: "This is what should be done, but it's a dream when you consider how difficult it is for an artist to get hosting platforms (Google Drive, Mega Drive, for example) to remove products uploaded by pirates"
This is likely because Content piracy is both international and legally a civil matter between the parties.
However google has the right to remove anything from their servers such as easily proven copyright claimed videos on youtube but again they make ad money from videos.
TBH they have less compelling business interest to quickly investigate claims of pirated 3D models on google drives etc. making them a lower priority.
There are no samples to remove from any libraries
CHKPT training data bases does not store any original images
just some code that represents its "knowledge "of how to create images based on LLM text prompts
:text=Diffusion%20models%20work%20by%20adding,order%20to%20produce%20new%20outputs." target="_blank"~https://arxiv.org/html/2312.14977v1#:~:text=Diffusion%20models%20work%20by%20adding,order%20to%20produce%20new%20outputs.
Quote: "This is what should be done, but it's a dream when you consider how difficult it is for an artist to get hosting platforms (Google Drive, Mega Drive, for example) to remove products uploaded by pirates"
This is likely because Content piracy is both international and legally a civil matter between the parties.
However google has the right to remove anything from their servers such as easily proven copyright claimed videos on youtube but again they make ad money from videos.
TBH they have less compelling business interest to quickly investigate claims of pirated 3D models on google drives etc. making them a lower priority.
REPLY
! REPORT
kwerkx
Karma: 6,151
Sun, Apr 13"CHKPT training data bases does not store any original images
just some code that represents its "knowledge "of how to create images based on LLM text prompts"
"Training data" is metadata trained on the samples (original images) in the training library.. this creates a "pre-trained" result (save) file that is delivered and used by the generative "AI" instead of the program having to redo the training step all over again.
The lawyers will have to argue if that training data file is it's own unique work or the a sum of the sample files and under what conditions the owners/creators of the samples have any legal rights.
NIL / Deep fake revenge porn / copyright infringement (Studio Ghibli) are the legal challenges that I hope will force the generative "AI" program owners to retrain their models after removing protected/prohibited content. Once the process is in place, I also hope it is easier to use and more effective than the "whack-a-mole" efforts @guy91600 describes with fighting content piracy.
You asked for realistic solutions, I think this is close to but not exactly your option #2.
just some code that represents its "knowledge "of how to create images based on LLM text prompts"
"Training data" is metadata trained on the samples (original images) in the training library.. this creates a "pre-trained" result (save) file that is delivered and used by the generative "AI" instead of the program having to redo the training step all over again.
The lawyers will have to argue if that training data file is it's own unique work or the a sum of the sample files and under what conditions the owners/creators of the samples have any legal rights.
NIL / Deep fake revenge porn / copyright infringement (Studio Ghibli) are the legal challenges that I hope will force the generative "AI" program owners to retrain their models after removing protected/prohibited content. Once the process is in place, I also hope it is easier to use and more effective than the "whack-a-mole" efforts @guy91600 describes with fighting content piracy.
You asked for realistic solutions, I think this is close to but not exactly your option #2.
I think Deep fakes should just be treayed as identity fraud and/or Defamation wether AI or traditional CG was used and depending on the context.
Quote: copyright infringement (Studio Ghibli) are the legal challenges
The studio Ghibli craze is not copyright infringement unless it involves a specific tradmarked character from that studio( Princess monoke etc)
You cannot copyright an Art style and the moment we establish the right to copyright a style the Big corporations like Disney or Amazon or Microsoft can buy up the rights to the major styles and lock them behind Pay/rent walls.
Quote: copyright infringement (Studio Ghibli) are the legal challenges
The studio Ghibli craze is not copyright infringement unless it involves a specific tradmarked character from that studio( Princess monoke etc)
You cannot copyright an Art style and the moment we establish the right to copyright a style the Big corporations like Disney or Amazon or Microsoft can buy up the rights to the major styles and lock them behind Pay/rent walls.
REPLY
! REPORT
ArtbyMel
Karma: 10,740
Mon, Apr 14According to the legalese I have read, it all depends on how the AI was trained, which would come out in discovery. If Studio Ghibli attorneys can prove that the AI was trained on the copyrighted material vs. the fanart, they have a good case. No one can sue based on copying style, but the main complaint with AI is about copyright violations based on AI training using actual material without permission or compensation. And that is perfectly understandable.
Quote: If Studio Ghibli attorneys can prove that the AI was trained on the copyrighted material vs. the fanart, they have a good case.
Why is their case any stronger than SamdoesArt
or any random artist on instagram?
or any random singer on spotify?
As of April 14 ,2025 no US court has ruled that using published visual/written audio content to train AI is illegal.
Also why would scraped FanArt be exempt?
My son (who is a career law enforcement officer)
Drew this Marvel/Disney "fan art" with colored pencils.
He obtained no permission from Marvel/Disney when he "scraped the internet" for reference images of marvels iteration of the "Thanos "character.
Studio Ghibli is not some sacred, special case
they will have to take their place in the legal que behind
everyone else who is trying to convince the courts that people should not be allowed to "train" AI. on content you published for the world to view.
Why is their case any stronger than SamdoesArt
or any random artist on instagram?
or any random singer on spotify?
As of April 14 ,2025 no US court has ruled that using published visual/written audio content to train AI is illegal.
Also why would scraped FanArt be exempt?
My son (who is a career law enforcement officer)
Drew this Marvel/Disney "fan art" with colored pencils.
He obtained no permission from Marvel/Disney when he "scraped the internet" for reference images of marvels iteration of the "Thanos "character.
Studio Ghibli is not some sacred, special case
they will have to take their place in the legal que behind
everyone else who is trying to convince the courts that people should not be allowed to "train" AI. on content you published for the world to view.
REPLY
! REPORT
Quite simply: generative AI can't challenge me 
....it's all too soulless, too highly polished and ultimately too stupid for me
I once put it this way (over on the dark side ('rotica)): if you show people the rude but soulful sketches of an Egon Schiele and, for comparison, high-gloss centerfolds, the MOB mainstream will always choose the bent ladies with the staples in the middle ... and I don't belong to the mainstream
....it's all too soulless, too highly polished and ultimately too stupid for me
I once put it this way (over on the dark side ('rotica)): if you show people the rude but soulful sketches of an Egon Schiele and, for comparison, high-gloss centerfolds, the MOB mainstream will always choose the bent ladies with the staples in the middle ... and I don't belong to the mainstream
REPLY
! REPORT
A program called nightshade is imbedded into your art and it poisons AI that try to use it.
https://nightshade.cs.uchicago.edu/whatis.html
https://nightshade.cs.uchicago.edu/whatis.html
REPLY
! REPORT
guy91600
Karma: 12,988
Tue, Apr 15Thank you for this information that every artist should know and use!
And here are probably my only two surviving
"traditional" paintings
Acrylic on stretched canvas
I Painted back when My son was a new born nearly 30 years ago.
Bygone Era.
REPLY
! REPORT
Daz Genesis 8 male and female
Original sci fi corridor model and animation, in Iclone ,by me
Rendered in Blender
No AI needed (this time)
REPLY
! REPORT
Getting back to the original thread subject
This Artist has suggested We try to convince the AI companies to license our images.
Would you license your work for AI training
if the payment was reasonable??
This Artist has suggested We try to convince the AI companies to license our images.
Would you license your work for AI training
if the payment was reasonable??
REPLY
! REPORT
FYI here is a link to my son's original instagram post of his hand drawn "Thanos "Fan Art from the year 2020
TWO YEARS before the public release of stability AI and Midjourney's image generators.
https://www.instagram.com/p/CH6e1MUBLTY/
TWO YEARS before the public release of stability AI and Midjourney's image generators.
https://www.instagram.com/p/CH6e1MUBLTY/
REPLY
! REPORT
