I request to move this thread to the Suggestion Box forum: https://www.renderosity.com/forums/12472/suggestion-box

Thank you.

Afrodite-Ohki posted at 7:28 AM Fri, 27 September 2024 - #4489787

pierremeu posted at 11:18 AM Thu, 26 September 2024 - #4489767

Because they borrow the model from someone else. Here you will create it yourself. I just want AI to give real texture to scenes. Nothing else. You don’t seems to understand what I envision. I would like to give my Poser man or woman a real photographic look and the same for nature. Turn your Poser scene into a real picture. But you have created the scene. 

To give your Poser man or woman a photographic look, that AI you want would have to  be trained on thousands of photographic images, currently all models that do this steal those images without their owners' consent.

This part has been repeated so often that nearly everyone believes it. First off, that is like saying you have stolen every bit of art you have ever seen. Even if your memory were perfect and you could reproduce every image you have ever seen with perfect fidelity, would you say that merely looking at something constitutes theft?
But go back to the AI models. There was a big deal about this last year when, while attempting to stoke fear over generative AI, a group were able to generate an image that matched an image in the training data almost exactly. It would be nearly impossible for this to happen by random chance. Its is most likely that the group knew which image sets the AI had been trained on. That the group knew a specific image was contained in more than one set of images, and that the image's content had been tagged. So, by inputting the tags that this specific image contained and then describing its contents and composition specifically they were in fact able to recreate a strikingly similar image. But they were aided in that they knew the image had been used in training, that it had been used more than once, and they knew the tags it contained when constructing their prompt. For some, those factors do not matter as they did prove they could retrieve an image by constructing the right prompt. 
But it's simply not possible for even a 10GB model to contain every detail of every image it has been trained on when that image set was 100s of GB or more.
Also, not all unlicensed use is theft. Fair use law for example allows the use of trademarked characters and even the likenesses of real people for the purpose of parody. Mad magazine can put E.T. or Captain American on their cover without infringing because no reasonable person would confuse the parody with the original work. There is also the possibility that an artist is creating works for their own enjoyment, and are not trying to profit from their creation in any way.
So while I have the same concerns about how AI might be used as others, whether to compete with artists trying to make a living or in creating works that could be mistaken for those of a specific known or unknown artist, I don't think that should be a reason to limit the training data.
There is also the issue of companies such as Getty now owning (via buying other companies' libraries) images that were produced before these models even existed. We think about protecting artists' rights as noble, but so much of what we think we are protecting no longer belongs to those artists. Artists who allowed Marvel to own their work could do nothing to prevent Disney from acquiring the rights to their work by buying Marvel for example. It is companies like Disney who now exist primarily to collect and monetize IP that are most threatened by generative AI, and they want you to fear it too. It threatens their efforts to monopolize content creation.

quote:
"This part has been repeated so often that nearly everyone believes it. First off, that is like saying you have stolen every bit of art you have ever seen. Even if your memory were perfect and you could reproduce every image you have ever seen with perfect fidelity, would you say that merely looking at something constitutes theft?"
Heard this so many times.
No, it is not the same.
Any human goes through individual personal experiences mixed with even genetic character traits, which is also part of the _inspiration_ .
I don't want a collection by software.
To put a human name as an artist under something, that a machine has spit out is dishonest.
AI "art" is not YOUR art, because legally you* don't own rights to what ever the software created. 
So I might steal this AI inspiration for my own pictures and take credits for them.

* Well, _you_ is not meant as a YOU- you, but for as a general unnamed addressing of users. ;-)

Versum posted at 7:47 PM Sat, 12 October 2024 - #4490243

Sure the argument , How nice the results are when you made a AI generated random Image, You might get a little attention, It might make you feel good for a moment, but deep inside " Listen to that little voice " it does not make you feel any better, as you know that any other person can do it as well. You do not get the feeling that you really painted Art you just told someone else how you want it to be painted. Not from your hand, at the end because it was not from your own hand it just cant be exactly the way you would of made it your self. You can compare it as if you went to a painter giving him a sheet of paper on how you would wish having the painting done trying to give him directives on how you wish having it using he's personal style. So who is the Artist and who will place the signature on the corner of the Paint ? Is it you ? or is it the painter who was told how to make the painting for you ?

You need to overthink, do you want to be the Artist, or do you want to be the one who tells the artist how to do it for you.

I want the results. When I was in Jr. High I began drawing on note cards to avoid having my drawings taken from me. I soon realized that drawing smaller meant I could draw the images faster. I stopped penciling first so the images wouldn't get smudged and found that saved even more time. When I started doing CG I'd have to let the computer render over night just to get a single image. Even now I browse or organize files while images are rendered. So part of the appeal of AI is the comparative speed of the image generation.
When I use AI to create images I get the enjoyment of seeing something that did not previously exist for the first time. I can choose the subject, the setting, the style and use other images to influence composition and the colors used. So I feel it's somewhere in-between. I don't have to spend hours refining every detail of each image, each one will be of similar quality. Photographers take multiple shots of a subject not knowing at the time exactly which ones they will choose to use or show. They may not even use the entire image, instead choosing to enlarge a single area of interest. Using AI is similar. You may generate dozens of images just to get one you like. But in choosing that image over the others you are doing essentially the same thing a photographer does when choosing a few images from among many.
Art should not be judged simply on the time, money, or effort put into it. Painting is time consuming, paintings are large and need space to dry. Canvas and paints are expensive. These are just a few of the reasons people like painting on tablets. In addition, digital images can be backed up, saved to the cloud or a thumb drive. But it wasn't long ago that people using Photoshop and similar programs were looked down on for not being "real" artists simply for using a novel, non-traditional, workflow.
I find it interesting to be having this conversation here, in a forum catering to users of a program best known for supplying a range of rigged figures and props for people to use in constructing virtual scenes. Why do we use pre-made figures? Because we don't all have the time or ability to make our own? Because we can't afford to hire models to pose for us? Doesn't using pre-made assets "rob" us of the satisfaction we'd get from making everything from scratch? Or does not having to model, UV map, and rig each figure/prop merely allow Poser users to focus on the things they DO like to do, such as scene composition, staging, lighting, atmosphere etc.?

Nevertrumper posted at 12:10 AM Mon, 14 October 2024 - #4490311

quote:
"This part has been repeated so often that nearly everyone believes it. First off, that is like saying you have stolen every bit of art you have ever seen. Even if your memory were perfect and you could reproduce every image you have ever seen with perfect fidelity, would you say that merely looking at something constitutes theft?"
Heard this so many times.
No, it is not the same.
Any human goes through individual personal experiences mixed with even genetic character traits, which is also part of the _inspiration_ .
I don't want a collection by software.
To put a human name as an artist under something, that a machine has spit out is dishonest.
AI "art" is not YOUR art, because legally you* don't own rights to what ever the software created. 
So I might steal this AI inspiration for my own pictures and take credits for them.

* Well, _you_ is not meant as a YOU- you, but for as a general unnamed addressing of users. ;-)

It's not like people press a generate button then print the first image they get and hang it in a gallery. They will generate dozens of images while noting the influence of words in the prompt over the end result. They'll change the order of words in the prompt, the strength or weight of terms and LORAs. They'll use different models. Ultimately they will select the images that best capture what they were trying to create initially. I don't see this as being too dissimilar to what a photographer does.
A photographer doesn't own the right to everything they may photograph. They don't own the likeness of a model or human subject. They don't own the likeness of the vehicles or buildings in the image. And of course any logos or trademarks in the image are already registered to someone else. Despite neither making nor owning any rights to the likenesses of the things they may choose to photograph the resulting image can still be copyrighted because of its unique composition and the arbitrary choices the photographer made from exposure to process. I feel AI generated works should be viewed in a similar light.
Maybe I've become too cynical but I have a hard time believing that the anti-AI tub-thumping, training prohibitions, and copyright exclusions specific to generative work are truly intended to protect individual artists and/or their works. Rather I feel, as I've stated, that it is the stock image libraries and multinationals in the IP business who are most threatened by the democratization of image and video generation. 
A gun that has been used to kill someone is a murder weapon. A similar gun that hasn't been isn't. I simply do not believe theft occurs at training, or even at generation. I believe theft occurs when an individual infringes on another's work with intent to deceive a client or infringe on the rights of a specific artist. And even then feel exceptions should remain for personal use, parody/satire, etc., just as they already do for non-generative works. 

moogal posted at 3:14 PM Sun, 13 October 2024 - #4490282

This part has been repeated so often that nearly everyone believes it. First off, that is like saying you have stolen every bit of art you have ever seen. Even if your memory were perfect and you could reproduce every image you have ever seen with perfect fidelity, would you say that merely looking at something constitutes theft?
But go back to the AI models. There was a big deal about this last year when, while attempting to stoke fear over generative AI, a group were able to generate an image that matched an image in the training data almost exactly. It would be nearly impossible for this to happen by random chance. Its is most likely that the group knew which image sets the AI had been trained on. That the group knew a specific image was contained in more than one set of images, and that the image's content had been tagged. So, by inputting the tags that this specific image contained and then describing its contents and composition specifically they were in fact able to recreate a strikingly similar image. But they were aided in that they knew the image had been used in training, that it had been used more than once, and they knew the tags it contained when constructing their prompt. For some, those factors do not matter as they did prove they could retrieve an image by constructing the right prompt. 
But it's simply not possible for even a 10GB model to contain every detail of every image it has been trained on when that image set was 100s of GB or more.
Also, not all unlicensed use is theft. Fair use law for example allows the use of trademarked characters and even the likenesses of real people for the purpose of parody. Mad magazine can put E.T. or Captain American on their cover without infringing because no reasonable person would confuse the parody with the original work. There is also the possibility that an artist is creating works for their own enjoyment, and are not trying to profit from their creation in any way.
So while I have the same concerns about how AI might be used as others, whether to compete with artists trying to make a living or in creating works that could be mistaken for those of a specific known or unknown artist, I don't think that should be a reason to limit the training data.
There is also the issue of companies such as Getty now owning (via buying other companies' libraries) images that were produced before these models even existed. We think about protecting artists' rights as noble, but so much of what we think we are protecting no longer belongs to those artists. Artists who allowed Marvel to own their work could do nothing to prevent Disney from acquiring the rights to their work by buying Marvel for example. It is companies like Disney who now exist primarily to collect and monetize IP that are most threatened by generative AI, and they want you to fear it too. It threatens their efforts to monopolize content creation.

That's a lot of words for "if I make the automated collage look sofisticated enough, I can pretend that machine learning is anything like a real human brain learning".

No, the machine learned photo collage is not like a real person using reference. Go off and stop pretending that people disagree with you because "we don't understand".

A person needs permission to use parts of someone else's work to do a photo collage and the machines should need permission to be fed from someone's work. It's not true learning, it's algorythm parsing. No matter how many times people decide to call it "artificial INTELLIGENCE", it's nowhere close to intelligence.

And this "But it's simply not possible for even a 10GB model to contain every detail of every image it has been trained on when that image set was 100s of GB or more" is not even an argument, nothing needs to contain every detail of something to be theft/plagiarism. Try using 5 seconds of a 10-minute-long copyright-protected song on your  Youtube video and see what will happen.

This thread has wandered far from the original question. What's being said now has been said many times. I'm going to lock this before tempers are lost and I have to start giving out warnings.