I went looking to see if a thread on this topic already existed. Please direct me appropriately if so.
AI Art Generation: Beneficial or Problematic?
Everyone and their mama has heard about AI art generators and the various arguments and opinions folks have about them by now. Stable Diffusion, Dream, Craiyon, even applications like Lensa that use AI to augment existing images and countless others are arriving by the day. They've quickly become a mainstay in almost every artistic community I'm apart of or adjacent to. I've heard people argue both in favor and against their use, especially when it comes to collecting a profit off what people generate, and now I want to hear it from ya'll.
Which side of the spectrum do you fall? Is AI a legitimate tool for people to use in artistic "creation", or do you feel more strongly about how it borrows from the artists these various engines "sample"? Do you think it's ethical for people to make money using these programs? Do you feel similar about AI text generators? Would you play a game, read a book, or consume other media that was made entirely from AI? Do you feel that artists should disclose whether or not AI was used in their works?
I'll sidestep the question and raise a point that should challenge users of AI art. Won't the data on which these AI tools are trained eventually become quite incestuous? If the AI developers scrape the web for art to feed into their algorithms, they're going to increasingly be feeding art generated by other AI tools, which could lead to a colossal garbage in/garbage out problem.
While I would hope you circle back around to my questions, I do find myself agreeing with you.
At the rate AI art is being generated, it's a matter of time if not years nee months before AI art overwhelms the kind exclusively created by human hands on the internet. In fact, we've already started to see what you're talking about happening in real time. So many of these engines (and so many of the creations I've seen flooding my own spheres) mimic each other in a way that's only put to shame by The Hills Have Eyes. Except with anime tiddies ¯\_(ツ)_/¯
However, I don't think this will slow the popularity of such applications and subsequent technology at all. People seem content with this level of creation as is. Before there was AI that generated 10 pictures of your waifu in variations of the same provocative pose, there was an artist with low creative/artistic standards producing almost the exact same content to the T. Lowest common denominator behaviour, if you will. I think the sort of people that consume that kind of art only care about that one thing: consumption.
I have no problem using AI art or AI writing tools. I think all AI productions should be labeled clearly as AI generated.
I have problems with AI scanning all of Picasso's art and then making a Picasso. It feels like it is a forgery. Or feeding it all of Mark Twain's works and having it write a Mark Twain short story. It feels like plagiarism. If, however, you clearly label it as an AI production and if it is based on a particular person's style, state that also, then it is less so. Just like attributing information in an essay to its source isn't plagiarism. It might not be great writing, but once it is attributed, then it isn't cheating.
When royalties and other income become involved, though, we're going to need something more than creative commons licensure.
I was going to post something similar to this, so I'm glad you did.
I'm not much of an artist, especially graphically, so I've started using AI to create character portraits and other art for my RPG life. Even if I were graphically inclined, I don't have the time to do all of the RPG stuff I do, pursue other hobbies and household chores and duties, and hold down a full time job.
I don't have a problem with that use of it. It saves me time and creates a product that I like.
My icon picture is AI generated, for example. I like it. It looks a lot like what I had in mind when I put in the prompt.
I mostly use StableDiffusionXL. It can be frustrating, though. It's ability to parse language lacks a bit. You can come up with some random stuff in the resulting image and key components omitted. If you ask it to use specific words, "Elf Ranger," for example, it often just produces gobbledygook.
It largely doesn't matter what people think of generative A.I.. The tool is here now and it's too powerful to put back in the box. Given ten years, generative A.I. will evolve dramatically, as will people's tools for interacting with and directing it. Artists either get in front of the trend and figure out how to coexist with generative A.I., use it to improve (or "improve") their work, or they're going to get left behind.
For what it's worth, I think we're losing a lot with the introduction of this tool. Replacing human artists with a fancy procedural generation tool reduces the human species and negatively impacts our shared global culture, but the tool is simply too prolific and powerful to put away now. Ordinary people with no artistic talent can now 'commission' bespoke art in multitudes, no longer having to wait weeks and pay hundreds of dollars for a single piece. They're not going to give that up, especially as the tools improve and as artists willing to take a pittance to 'fix' A.I.-generated pieces rather than produce from-scratch original work become easier to find.
It sucks. We're voluntarily throwing away an incredibly valuable skillset that's been part of our species' history since before homo sapiens was a species, and I absolutely hate it. But we'd all better get used to it because there's scads of money to be made in keeping it, and we all know how that fight ends.
I just looked at 4 different AI Art sites. I wanted a female feline humanoid barbarian swinging a two handed sword over her head. One was abysmal. Too blurry to make much out. Another literally pasted cat heads to a body. Another did the same just a little less obvious. It did have a tail for the left arm with a sword on it and the arm as the tail. Haha. The last had the best overall art but the thigh looked like it was a calf and the right foot that was lifted in the air had a double paw thing going on. The left hand was also backwards pressed against the sword. Needless to say, the AI art isn't up to snuff yet by quite a bit.
My experience with AI art is that you can't get too complicated or expect too much. It takes a lot of iterations to get something suitable for your needs. I was real happy the steel protector it produced, but my request was pretty straightforward. When I asked for the gnome artificer to be smoking a pipe, it frequently came disarticulated and the bowl sticking out of his head.
AI is going to created, trained, and maintained by someone. If the track record of Microsoft, IBM, Google, or Apple is any indication, we needn't fear the overlords just yet.
AI technology and how to use it or limit its use is a major debate and interestingly those making policy generally have little experience using AI. Much of the AI art I se is of various quality. /it is because of the quality of the program and the person inputting the parametars. I’m a fan of good AI art and like other artistic platforms, I don’t see anything wrong making money through AI art. The legal and ethical gay zone is using the likeness of real individuals without permission. I am an amateur photographer, and if I take a really great photo and someone is in it and clearly identifiable, I would need to get their permission to publish or sell their image. I think similar rules/laws need to be set in place for AI art as I regularly see celebrity and noncelebrity images used as the foundation with little alteration of the person’s features.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
I went looking to see if a thread on this topic already existed. Please direct me appropriately if so.
AI Art Generation: Beneficial or Problematic?
Everyone and their mama has heard about AI art generators and the various arguments and opinions folks have about them by now. Stable Diffusion, Dream, Craiyon, even applications like Lensa that use AI to augment existing images and countless others are arriving by the day. They've quickly become a mainstay in almost every artistic community I'm apart of or adjacent to. I've heard people argue both in favor and against their use, especially when it comes to collecting a profit off what people generate, and now I want to hear it from ya'll.
Which side of the spectrum do you fall? Is AI a legitimate tool for people to use in artistic "creation", or do you feel more strongly about how it borrows from the artists these various engines "sample"? Do you think it's ethical for people to make money using these programs? Do you feel similar about AI text generators? Would you play a game, read a book, or consume other media that was made entirely from AI? Do you feel that artists should disclose whether or not AI was used in their works?
Why or why not?
I'll sidestep the question and raise a point that should challenge users of AI art. Won't the data on which these AI tools are trained eventually become quite incestuous? If the AI developers scrape the web for art to feed into their algorithms, they're going to increasingly be feeding art generated by other AI tools, which could lead to a colossal garbage in/garbage out problem.
While I would hope you circle back around to my questions, I do find myself agreeing with you.
At the rate AI art is being generated, it's a matter of time if not years nee months before AI art overwhelms the kind exclusively created by human hands on the internet. In fact, we've already started to see what you're talking about happening in real time. So many of these engines (and so many of the creations I've seen flooding my own spheres) mimic each other in a way that's only put to shame by The Hills Have Eyes. Except with anime tiddies ¯\_(ツ)_/¯
However, I don't think this will slow the popularity of such applications and subsequent technology at all. People seem content with this level of creation as is. Before there was AI that generated 10 pictures of your waifu in variations of the same provocative pose, there was an artist with low creative/artistic standards producing almost the exact same content to the T. Lowest common denominator behaviour, if you will. I think the sort of people that consume that kind of art only care about that one thing: consumption.
Howdy y'all!
I have no problem using AI art or AI writing tools. I think all AI productions should be labeled clearly as AI generated.
I have problems with AI scanning all of Picasso's art and then making a Picasso. It feels like it is a forgery. Or feeding it all of Mark Twain's works and having it write a Mark Twain short story. It feels like plagiarism. If, however, you clearly label it as an AI production and if it is based on a particular person's style, state that also, then it is less so. Just like attributing information in an essay to its source isn't plagiarism. It might not be great writing, but once it is attributed, then it isn't cheating.
When royalties and other income become involved, though, we're going to need something more than creative commons licensure.
Huzzah!
Jack
Howdy Zweit!
I was going to post something similar to this, so I'm glad you did.
I'm not much of an artist, especially graphically, so I've started using AI to create character portraits and other art for my RPG life. Even if I were graphically inclined, I don't have the time to do all of the RPG stuff I do, pursue other hobbies and household chores and duties, and hold down a full time job.
I don't have a problem with that use of it. It saves me time and creates a product that I like.
My icon picture is AI generated, for example. I like it. It looks a lot like what I had in mind when I put in the prompt.
I mostly use StableDiffusionXL. It can be frustrating, though. It's ability to parse language lacks a bit. You can come up with some random stuff in the resulting image and key components omitted. If you ask it to use specific words, "Elf Ranger," for example, it often just produces gobbledygook.
I'd love to hear what other people think of it.
Huzzah!
Jack
It largely doesn't matter what people think of generative A.I.. The tool is here now and it's too powerful to put back in the box. Given ten years, generative A.I. will evolve dramatically, as will people's tools for interacting with and directing it. Artists either get in front of the trend and figure out how to coexist with generative A.I., use it to improve (or "improve") their work, or they're going to get left behind.
For what it's worth, I think we're losing a lot with the introduction of this tool. Replacing human artists with a fancy procedural generation tool reduces the human species and negatively impacts our shared global culture, but the tool is simply too prolific and powerful to put away now. Ordinary people with no artistic talent can now 'commission' bespoke art in multitudes, no longer having to wait weeks and pay hundreds of dollars for a single piece. They're not going to give that up, especially as the tools improve and as artists willing to take a pittance to 'fix' A.I.-generated pieces rather than produce from-scratch original work become easier to find.
It sucks. We're voluntarily throwing away an incredibly valuable skillset that's been part of our species' history since before homo sapiens was a species, and I absolutely hate it. But we'd all better get used to it because there's scads of money to be made in keeping it, and we all know how that fight ends.
Why you shouldn't start ANOTHER thread about DDB not giving away free redeems on your hardcopy book purchases.
Thinking of starting ANOTHER thread asking why Epic Boons haven't been implemented? Read this first to learn why you shouldn't!
I just looked at 4 different AI Art sites. I wanted a female feline humanoid barbarian swinging a two handed sword over her head. One was abysmal. Too blurry to make much out. Another literally pasted cat heads to a body. Another did the same just a little less obvious. It did have a tail for the left arm with a sword on it and the arm as the tail. Haha. The last had the best overall art but the thigh looked like it was a calf and the right foot that was lifted in the air had a double paw thing going on. The left hand was also backwards pressed against the sword. Needless to say, the AI art isn't up to snuff yet by quite a bit.
Howdy Nazarethll!
My experience with AI art is that you can't get too complicated or expect too much. It takes a lot of iterations to get something suitable for your needs. I was real happy the steel protector it produced, but my request was pretty straightforward. When I asked for the gnome artificer to be smoking a pipe, it frequently came disarticulated and the bowl sticking out of his head.
AI is going to created, trained, and maintained by someone. If the track record of Microsoft, IBM, Google, or Apple is any indication, we needn't fear the overlords just yet.
Huzzah!
Jack
AI technology and how to use it or limit its use is a major debate and interestingly those making policy generally have little experience using AI. Much of the AI art I se is of various quality. /it is because of the quality of the program and the person inputting the parametars. I’m a fan of good AI art and like other artistic platforms, I don’t see anything wrong making money through AI art. The legal and ethical gay zone is using the likeness of real individuals without permission. I am an amateur photographer, and if I take a really great photo and someone is in it and clearly identifiable, I would need to get their permission to publish or sell their image. I think similar rules/laws need to be set in place for AI art as I regularly see celebrity and noncelebrity images used as the foundation with little alteration of the person’s features.