Personally, what makes me cringe the most about AI art these days is how many people on social media get so much attention/engagement/praise for sharing their midjourney and whatnot prompts like they have anything to do with that. Meanwhile, thousands of great artists/photographers/filmmakers whose works these AI platforms "learn" from are left unnoticed due to ever so ruthless social media algorithms.
While I agree they have a lot of attention for that now, it's not going to last very long as it becomes super oversaturated. Eventually, you'll have to back it up with a lot more effort and something unique. But I would say that's how any industry works. In the music industry, most of the best musicians never get the recognition they deserve because they don't know how to market themselves. While you could say that the people who use midjourney prompts "don't have to do anything." They still came up with a creative idea that resulted in those videos becoming popular for the time being. Which requires a lot of work keeping track of trends to make content on. Not trying to take anything away from all the artists, photographers, and filmmakers out there who go unnoticed, but at the end of the day the platforms like tiktok or UA-cam are a specific style of content and sometimes isn't necessarily the best place for artists if they don't want to edit for their audience. As much as UA-cam's algorithm's annoy me at times, it's fair. If people aren't interested in the content, they won't keep showing it. But please, don't take this as an attack on your opinion. Because it's completely valid. We prioritize the wrong things at times when it comes to who we support.
@@FrameVoyager or companies can start dividing ai generated content from non-generated content. So that those who want ai stuff can go there and those who don't can go the other way. I know it's not that simple and that bad faith actors will do what they always do, but at the very least it's something.
I can't tell you how many viral videos there are of AI art compilations and people mindlessly saying "beautiful" Although who is to say they're not view botted? But... Ugh.
@@FrameVoyager There is one aspect of this that seems like it is being over looked. The fact some companies are charging users to us these AI's or profiting off advertisements that sponsor the available prompt, but the artists who's content was used in training those AI's don't see anything form those profits. IF everything were made open source & public domain, & these companies were unable to profit off their AI's & the AI's art, then one could argue that no one gained anything from the whole AI process, but that simply isn't happening. There are hundreds of AI art generation apps for phones that WERE trained off of copyrighted material that appeared in public art databases, & used without the artists permission or their compensation. Yes, there is an inevitability that this was eventually going to occur with AI development; however, if the AI in question was indeed trained with an artists work, and with out compensating the artist, & the company then makes money off that AIs creations then the original artists absolutely should be included in the distribution of those profits, as their art was used for commercial purposes without first gaining & paying for their permission to use it for those purposes. The art generated after the fact also needs to remain public domain as it isn't human created, and any modifications to the art would need to pass the collage modification laws in order to be considered art created by an individual, which a text prompt can not possibly be considered part of. The text prompt it self could potentially be copyrighted text, but the resulting image can't, because it isn't human generated, and thus no human can hold copyright on that result without, again, passing the collage modification requirement laws. AKA it can't be easily identified as coming mostly from the original source image alone. This last bit is a very grey area that hasn't been set in stone in courts, because, up till now, collage of this type & magnitude was a human only thing that took a lot of time, effort & work to achieve. Typically if the resulting collage was different enough it was seen as a unique creation. But let us not confuse a a patron requesting a type of picture (aka the prompter) with the artists creations aka the AI's result. Unless work for hire the patron doesn't own any part of the resulting materials copyright, and at the point where an AI's art might be considered work for hire is past the point at which the artist work contribution and compensation that would have trained the AI would have come into play. Until the compensation aspect has been resolved properly then the AI's work should not legally be considered capable of being work for higher, but at best public domain.
Having achieved a Masters In Fine Art I have two thoughts A. This is the dream come true for modernists. All media available instantly to manipulate infinatly. B. Traditional fine art just became the most valuable manmade commodity on the planet.
I agree! I think they are going to have to add some regulations around all of it at some point. But it's interesting to see what it could be used for. I still don't think it will take away from traditional fine art like you said. It's going to be a big commodity.
@@FrameVoyager oops. I didnt mean to say it's taking away from fine art as it actually makes unique creation even more valuable! The AI dosnt copy itself that well yet...so I think original artists can copyright their works with tools likenft's and that will go along way to protect their creations by forcing the algoritm to prove work to copy it. Like a water mark for ai.
Yeah idk about this one man. I like the content but ethically it’s a swing and a miss. I don’t think that there is a credible way to run the “AI art is good or bad” discussion while being sponsored BY an AI art company. I’m glad that you popped the copyright question, but I also don’t like their response. “It’s inevitable, just get over it“ is quite frankly a low effort BS argument. People get their identities stolen as more and more of our information winds up online. Doesn’t mean we shouldn’t be working hard to limit the damage just because it’s gonna happen whether we like it or not. I appreciate your transparency on the matter, and disclosure of research though. Those are all great moves. In the future, I just would recommend not being sponsored by somebody who has a vested interest in the polarizing side of your argument. Especially when it’s this culturally relevant or charged.
Totally get it. And while I understood it would come across that way it did open up some insights into how ai companies are thinking. The whole script was written before we were approached by them as well and we inserted the interview throughout the video. But it's also through that we were able to get the interview which I felt we needed for this video. We tried to balance both sides of the argument and while we were sponsored by them, there was no video review process from them on this video. It was more of a showing that this was available for free, unlike MidJourney, to use and explore if you wanted to see what it was about. This was a pretty unique situation for this piece of content and we probably won't do anything like this again. Appreciate your honesty!
Thanks for the reply and clarification - I understand what you mean about getting the interview. Feels a little better knowing that, and it might help to include that info in the future should you wind up in a similar situation.
@@ColePerrine You're right! Probably could have done with a bit of clarification on this. I know personally, I'm big on being as neutral as possible on our videos and kind of letting the viewer decide as much as I can. We would never take a sponsorship that would force us to say something we don't want to say or keep us from saying something. But I also totally understand how that can come across to a viewer. Still trying to figure out how to approach all of this properly as the sponsors do help to conduct better research, but conversations like this help me think through what we should do in future videos. So I really do appreciate you taking the time to voice that concern and perspective.
If I went into Photoshop and hit "Render Clouds" and sold that image, does that mean it would be public domain because I literally did nothing? I see AI as a tool for creators, even if the effort was minimal, it's still comes down to the creator, the one to initiate the process. Some people get bent out of shape because of the amount of effort to create something was in some way reduced. Art is in itself subjective so arguing what is "good" art and "bad" art should not be the debate based on the effort that went into it. If I gave a paint brush to an elephant and had it paint me a picture and sold that picture even though I did nothing but supply the paint and the brush, is that legal for me to do? Do I own the picture or does the elephant? In that case, it's a collaborative effort. AI prompts require human input to generate art and as such, has a human element to it. At the very least, the developers who wrote the software could be the artists, if no one else. AI as we think of it is nothing more than software, like Photoshop, a tool for the creative person.
Totally agree! I think the only thing to really figure out is the copyright side of all the photos and files being used to train the systems. That's what really needs to be fixed here early on
@@FrameVoyager agreed. The funny thing is, if I took an art class to learn how to draw or shoot photography, what would I be shown how to draw or to shoot photography? Would I not be given examples of works. In fact, the instructor may even ask me to paint a familiar photograph or to replicate an existing work to train me? What’s the difference? I guess the courts will decide.
@@CNC-Time-Lapse If you were in art class, you'll start with how to hold your pencil sketching objects. Then maybe human pose with a real model. In the class I attended, the school prepare a premade artwork that is done by the instructor and asked to copy it. Of course might be different for other school since the one I attended was taught by Japanese teacher. And they are very careful of legal sides. We were only taught to copy or follow the material that he prepared. And that is only at the near end of class only after real life sketching and theories.
This small detail, that something must be made by a human in order to have copyright, is so important and is being overlooked by so many people. This video should have way more views.
To be honest, I think this could empower small businessess. It's like typography which only people with a lot of money could afford. But nowadays changing a typeface is built into word. With AI Art you don't need expensive marketing consultants to make an Instagram/Facebook ad. And regarding to art, I think artists are the og influencers. You don't really follow the art you follow the personality behind it. We shouldn't worry about AI art, bots and deep-fakes are the real danger!
Yep! Most of art is the personality and story behind it. Eventually the shine of the AI art that everyone is using will die down and people will actually have to figure out how to use it well.
I don’t necessarily see a problem with you using it for your videos because Photoshop has existed long before which gives you the ability and flexibility to doctor images, albeit, from a person utilizing software and a computer. Where it gets iffy is when it’s used to doctor falsified imagery/news. What’s more scary is the technology behind deep fakes where people’s faces can be put onto another’s body. It’s quite a rabbit hole, and I don’t see it being de tangled anytime soon. It’s a complex and interesting conundrum 🤔
Totally agree. Like the one article I shared with them using ai photos in Syria/Turkey to scam people into donating. That's why we created our own policy so you can always check and ask if we are using it.
Love this, my creative partner and I have been talking about how we could best use AI to make what we do more efficient/increase our output/ect. Basically what you’re taking about, regardless if we like it or not, we should probably figure out how can use it to our advantage.
Pretty much. It's coming and the sooner you figure out how it works, the faster you can integrate it. With that being said, you can still choose to use it as ethically as possible and support rights for photographers
I think AI art (or any AI generated content) is more or less like saying to someone, I want this (whatever), could you create it for me as long as I can modify it how I want? Its not yours in that you didn't create it, you just asked for it and modified it to your liking. It doesn't take away genuine skills of being able to create artwork without it, but I do think it possibly could take away creativity when technology can mostly do it for you. Sure you could argue there have been tools around for years like Photoshop that allows you do manipulate images, but that is the person doing it, with skill, not a computer generating an image that you don't have to do much to make it how you want. I also think that AI tech should not be exploiting other people's work as a basis for its learning, unless they give permission first as that to me it is a bit like forgery (when it is AI art) or stealing. Its unethical and I can see more debate and possibly laws introduced if AI technology like this keeps advancing and becoming more widely used.
I guess it depends on what you do with it? I mean I can point and shoot with an ARRI camera and get good footage but if I have no clue how to process it or edit it into something good, does it even matter how it was shot in the first place? I think a similar philosophy could be applied to AI art. Eventually, the market will become saturated with it and only the people who understand how to take the initial image or prompts and create something else with it will end up with anything to show for it. For me, it's a tool for a small piece of what I do and is not the cornerstone of the content I make. But I totally agree with you on the ethical side of it. Stable Diffusion just got a class action lawsuit, so while it's going to be an interesting legal battle, I think we will start to see some rules and regulations for this being put in place. Because AI is the next wave of technology whether we want it or not, companies have invested too much into it. Thanks for the great comment!
The side broheim has about *needing* to work in conjunction with the AI, because otherwise you'd become uncompetitive is not a very POG point of view, essentially. AI creators want you to forcefully embrace it as a consolation gift for having everyone's work being used.
I totally agree that AI datasets should have regulations and pay royalties for work they've used. I'm glad to see the class action lawsuit against stable diffusion as that should hopefully start setting some precedents for the industry going forward. Better now than way later down the line. But I wouldn't say you would uncompetitive, but I think it's a useful tool to understand even if it's not ready to be fully utilized yet. But I totally get the sentiment, a lot of people feel the way you do about it.
@@FrameVoyager The problem I see with any sort of royalty payout is that due to the large number of images involved, even if there is a large pool of money set aside, the splits could be so small that it doesn’t really do that much financially for the artists and perhaps undercuts their argument in the process. Additionally, if we’re dealing with a generated work, if hundreds or thousands of artists are pointing to the same parts of the work and claiming “it copied me”, that also undercuts their argument by reducing the distinctness of their work. As you can probably guess, I’m fairly skeptical of the lawsuit against AI art generators. I’m not completely unsympathetic to artists, and believe that they deserve credit and compensation for more direct usage of their work, but when you try to apply that to a statistical average/median of a large number of similar elements (which is kind of what these models are, based on my technical understanding), the logistics and arguments don’t hold up terribly well to me. I wouldn’t call myself a great fan of how the data was collected for these models, but I think they have a viable argument from a legal and to some extent even moral standpoint. We’ll see how it plays out in court.
The "big corporations taking away our art" argument falls apart pretty quickly when you consider that in a few years, users will be able to train these things themselves. Right now training via back propagation is extremely expensive and tagging requires a lot of human input, but this could change with for example, using the Forward-Forward method on dedicated analog ML hardware or other improvements, this is a rapidly changing field. Once people get their models trained themselves with an AI scraping the web and tagging the photos then the moral distinction of you learning something vs your computer doing it will be non existent.
I don't think it is complicated at all. It is fairly simple people should not be exploited, there is no ethical basis for exploiting people. So to train AI algorithms on artists work without compensation is unethical and I think there is a legal case to argue that it is theft. If you prompt an AI and it uses my artwork to generate something for someone else, I should be compensated, particularly when these companies have monetized the process. You have to pay for DALLE and Midjourney so they are making money and compensating none of the artist in their vast reserve of art. It is unethical. If I could input my own art, if I could do a drawing or film something and input it into the ai and say make the drawing photo real...that is totally different. If I render a building or object and I say take this render and place it in this picture of x and the AI automates the process that is totally a different thing. What is is doing now is in no way shape or form ethical. the artist in their banks should be paid and every time their art is used to generate a prompt they should be compensated. The computer should know which image or set of images it uses to generate your prompt and some monetary compensation should go to the artist. Also artist should have the option to opt out of this, like any random company can come along and sweep up your images into its training model with no notice to you and again, no compensation. I just don't get the confusion about this idea of having a monetized product that is generating revenue with no compensation to the artist whose work you are using to generate that revenue. It is plainly exploitation and it is unethical. All you have to do with this model is share in the revenue, make it so the AI keeps track of what it is using to generate the prompt and pay those artist...then what you'll have is everyone rushing to upload their work to your training module, and signing on to partnerships with these companies.
Totally agree. And honestly probably where this is all headed. But I do think the legal question is going to be a very murky one. Though I do get the sense from some of the ai companies that they will want to make some sort of agreement. Probably good for short term viability and to clear these specific ethical questions
@@FrameVoyager The thing is that the explosion of AI art generator companies (using Stable Diffusion AI motor as their core) that it is happening today has just one reason behind, making money before the party is off. They already know that they gonna end having legal troubles in the future (well for Stable Diffusion is a present thing) so they need to squeeze the AI til the last drop before the law forced them paying for each copyrighted image their AI use or, worst, to only relay in a dataset wich only contains non-copyrighted pictures, making their generation of images a couple steps back to current technology. The thing is that only a few current pro artist (probably less than 20%) gonna end using that technology, at least for a time, because as it is currently developed can't offer much to a pro artist, unless the style of the artist is very in pair with what Stable Diffusion or Dall-e can already produce, that apart from photorealism, the AI models usually have a very narrow spectrum of styles even if they were trained and feed with the majority of the artist from this world, even my artwork is in in the database used with Stable Diffusion and I can assure you that the AI can't produce any image near to what I can produce, it is useless for me unless I use as a reference in the same way I use Google or Pinteres (If I use them, coz usually I not). At the end is pure late capitalism, same shit as cryptocurrency and NFTs, be in the party first, make as much money as you can, and run away before things end badly.
@@Akenoomokoto Totally agree with you on the stable diffusion thing. And they just announced a class action lawsuit against that dataset. Which I am totally for. This is the time to figure out the legality issues and make sure that people get paid fairly for their work being used. I think a lot of this will have to be worked out because AI tech is going to be here to stay. It's not where it needs to be yet for most professional use cases, but in many regards it's still really in its infancy and every time a prompt is made it gets better and gains more information. I think it's going to be the future in some way. We need regulations and copright laws for it for sure, but I think this is different than crypto of NFT's. But totally get your point
@@FrameVoyager well popular NFTs like "bored ape" were made with AI generators trained by their authors. NFTs make some how popular the use of this kind of tools to make profit from the and actually is precisely in NFTs (and some websites like Society 6) where the AI bros are trying to make money (there is hundreds of videos here in UA-cam explaining how yo monetize AI images using NFTs technology and some other website, the majority of them promise impossible amount of money...). I'm not saying that AI art generators are the same thing, I mean, I said it, but I don't mean exactly that, what I mean is that this kind of tools and technologies are under regulated, have the same spread pattern, have the same late capitalism ideology behind, and have the same end users, at least in its initial shape, people who do not care about ethics (because NFTs are... Well also not very ethic) and the main interest is make some short of profit from them. I'm already aware that in the case of AI art generator software the target has a wider spectrum, there is people that just want to use for fun (but I don't see they paying for it) and also there is plenty of content creators that could use it for their video thumbnails, their Channel content and stuff like that, but... That's just collateral targets, or at least that's what I'm seeing now a days.
I'm all for innovation, but if the images used for generating the AI models weren't licensed by the original artists, don't the AI companies need to pay the artists for corporate use of their work? Not in terms of 'per image created artists are owed money' but in terms of the original use of the image to train the AI?
That's why there is a class action lawsuit against stable diffusion currently. It's more than likely going to set that very precedent. But not every company is doing that. It's kind of the wild west right now and hopefully we start getting some sheriffs in town haha
I certainly hope it reaches a point of agreement for all sides. I imagine in a few years Bing images will be a very different space both in search and legal terms (because of course we've all moved to Bing with Chat GPT 😉) Great coverage of this by the way, loved reliving the Gorrilla lawsuit 😂
@@astrobotnautics5291 Same! But gosh if Bing's AI chat isn't going to be the one that ends us all I don't know what is haha. That thing is scary. And appreciate it!
Personally I would only be offended if my original imagery is being used to create profit somewhere thereafter. Artists should be then compensated the same way they do with stock imagery. My work is being used in reproduction for profit? I want royalties. It’s one thing to talk about art and it’s another thing to talk business. Blue Willow obviously believes it’s all for the best it’s a business too. Can I ask an AI to attempt to draw up code for a program that takes their IP and uses it for whatever I deem “good for humanity” ? I don’t doubt the possibilities but it seems like a lot is being done in grey areas with little regard on ethics
Exactly. That's why I'm really hoping the class action lawsuit being brought against stable diffusion helps to start setting some precedent and rules and regulations. It's great tech, but it's the wild west right now
AI art sucks because they never asked artist consent for widespread usage of their data. Not a single penny of compensation, while the AI Art companies reap the profits from peoples hard work. Its a despicable practice that will lead to countless of lost jobs and its ALREADY happening in many studios. Countless of audiences & workers are creeped out by un-ethical business practice......because you say "but you can use it to further your work"....i tell you this. if you are artist working in the creative industries that LOVE making what they make, love the creative workflow you have. By saying that, you are essentially telling us to ditch the thing that we so desperately love & crave to do. By saying that you are UNDERVALUING the worth of what artists and fellow creatives do. I don't WANT AI to improve my workflow. I dont NEED AI to do that. I can improve my workflow with my own brain power. I am PROUD of my hard work and journey, but AI threatens to wipe that out. Further to this, creatives working in the industry DESERVE legal protections and regulations against AI companies scraping data without compensating those who made the data. It should be Opt Out by default, and Opt In is optional.....but the current model is the exact opposite. Also, of course the CEO will side with his AI business practice. OF COURSE he would, its making him HUGE profits. This is a biased take.
Personally, what makes me cringe the most about AI art these days is how many people on social media get so much attention/engagement/praise for sharing their midjourney and whatnot prompts like they have anything to do with that. Meanwhile, thousands of great artists/photographers/filmmakers whose works these AI platforms "learn" from are left unnoticed due to ever so ruthless social media algorithms.
While I agree they have a lot of attention for that now, it's not going to last very long as it becomes super oversaturated. Eventually, you'll have to back it up with a lot more effort and something unique.
But I would say that's how any industry works. In the music industry, most of the best musicians never get the recognition they deserve because they don't know how to market themselves. While you could say that the people who use midjourney prompts "don't have to do anything." They still came up with a creative idea that resulted in those videos becoming popular for the time being. Which requires a lot of work keeping track of trends to make content on.
Not trying to take anything away from all the artists, photographers, and filmmakers out there who go unnoticed, but at the end of the day the platforms like tiktok or UA-cam are a specific style of content and sometimes isn't necessarily the best place for artists if they don't want to edit for their audience. As much as UA-cam's algorithm's annoy me at times, it's fair. If people aren't interested in the content, they won't keep showing it.
But please, don't take this as an attack on your opinion. Because it's completely valid. We prioritize the wrong things at times when it comes to who we support.
@@FrameVoyager or companies can start dividing ai generated content from non-generated content. So that those who want ai stuff can go there and those who don't can go the other way. I know it's not that simple and that bad faith actors will do what they always do, but at the very least it's something.
I can't tell you how many viral videos there are of AI art compilations and people mindlessly saying "beautiful"
Although who is to say they're not view botted? But... Ugh.
It's similar to how cringe it is to call recorded gameplay a "derivative work"
@@FrameVoyager There is one aspect of this that seems like it is being over looked. The fact some companies are charging users to us these AI's or profiting off advertisements that sponsor the available prompt, but the artists who's content was used in training those AI's don't see anything form those profits. IF everything were made open source & public domain, & these companies were unable to profit off their AI's & the AI's art, then one could argue that no one gained anything from the whole AI process, but that simply isn't happening. There are hundreds of AI art generation apps for phones that WERE trained off of copyrighted material that appeared in public art databases, & used without the artists permission or their compensation. Yes, there is an inevitability that this was eventually going to occur with AI development; however, if the AI in question was indeed trained with an artists work, and with out compensating the artist, & the company then makes money off that AIs creations then the original artists absolutely should be included in the distribution of those profits, as their art was used for commercial purposes without first gaining & paying for their permission to use it for those purposes. The art generated after the fact also needs to remain public domain as it isn't human created, and any modifications to the art would need to pass the collage modification laws in order to be considered art created by an individual, which a text prompt can not possibly be considered part of. The text prompt it self could potentially be copyrighted text, but the resulting image can't, because it isn't human generated, and thus no human can hold copyright on that result without, again, passing the collage modification requirement laws. AKA it can't be easily identified as coming mostly from the original source image alone. This last bit is a very grey area that hasn't been set in stone in courts, because, up till now, collage of this type & magnitude was a human only thing that took a lot of time, effort & work to achieve. Typically if the resulting collage was different enough it was seen as a unique creation. But let us not confuse a a patron requesting a type of picture (aka the prompter) with the artists creations aka the AI's result. Unless work for hire the patron doesn't own any part of the resulting materials copyright, and at the point where an AI's art might be considered work for hire is past the point at which the artist work contribution and compensation that would have trained the AI would have come into play. Until the compensation aspect has been resolved properly then the AI's work should not legally be considered capable of being work for higher, but at best public domain.
Having achieved a Masters In Fine Art I have two thoughts
A. This is the dream come true for modernists. All media available instantly to manipulate infinatly.
B. Traditional fine art just became the most valuable manmade commodity on the planet.
I agree! I think they are going to have to add some regulations around all of it at some point. But it's interesting to see what it could be used for. I still don't think it will take away from traditional fine art like you said. It's going to be a big commodity.
@@FrameVoyager oops. I didnt mean to say it's taking away from fine art as it actually makes unique creation even more valuable! The AI dosnt copy itself that well yet...so I think original artists can copyright their works with tools likenft's and that will go along way to protect their creations by forcing the algoritm to prove work to copy it. Like a water mark for ai.
Yeah idk about this one man. I like the content but ethically it’s a swing and a miss. I don’t think that there is a credible way to run the “AI art is good or bad” discussion while being sponsored BY an AI art company.
I’m glad that you popped the copyright question, but I also don’t like their response. “It’s inevitable, just get over it“ is quite frankly a low effort BS argument. People get their identities stolen as more and more of our information winds up online. Doesn’t mean we shouldn’t be working hard to limit the damage just because it’s gonna happen whether we like it or not.
I appreciate your transparency on the matter, and disclosure of research though. Those are all great moves. In the future, I just would recommend not being sponsored by somebody who has a vested interest in the polarizing side of your argument. Especially when it’s this culturally relevant or charged.
Totally get it.
And while I understood it would come across that way it did open up some insights into how ai companies are thinking. The whole script was written before we were approached by them as well and we inserted the interview throughout the video. But it's also through that we were able to get the interview which I felt we needed for this video.
We tried to balance both sides of the argument and while we were sponsored by them, there was no video review process from them on this video. It was more of a showing that this was available for free, unlike MidJourney, to use and explore if you wanted to see what it was about.
This was a pretty unique situation for this piece of content and we probably won't do anything like this again. Appreciate your honesty!
Thanks for the reply and clarification - I understand what you mean about getting the interview. Feels a little better knowing that, and it might help to include that info in the future should you wind up in a similar situation.
@@ColePerrine You're right! Probably could have done with a bit of clarification on this. I know personally, I'm big on being as neutral as possible on our videos and kind of letting the viewer decide as much as I can. We would never take a sponsorship that would force us to say something we don't want to say or keep us from saying something. But I also totally understand how that can come across to a viewer.
Still trying to figure out how to approach all of this properly as the sponsors do help to conduct better research, but conversations like this help me think through what we should do in future videos. So I really do appreciate you taking the time to voice that concern and perspective.
If I went into Photoshop and hit "Render Clouds" and sold that image, does that mean it would be public domain because I literally did nothing? I see AI as a tool for creators, even if the effort was minimal, it's still comes down to the creator, the one to initiate the process.
Some people get bent out of shape because of the amount of effort to create something was in some way reduced. Art is in itself subjective so arguing what is "good" art and "bad" art should not be the debate based on the effort that went into it. If I gave a paint brush to an elephant and had it paint me a picture and sold that picture even though I did nothing but supply the paint and the brush, is that legal for me to do? Do I own the picture or does the elephant? In that case, it's a collaborative effort.
AI prompts require human input to generate art and as such, has a human element to it. At the very least, the developers who wrote the software could be the artists, if no one else. AI as we think of it is nothing more than software, like Photoshop, a tool for the creative person.
Totally agree! I think the only thing to really figure out is the copyright side of all the photos and files being used to train the systems. That's what really needs to be fixed here early on
@@FrameVoyager agreed. The funny thing is, if I took an art class to learn how to draw or shoot photography, what would I be shown how to draw or to shoot photography? Would I not be given examples of works. In fact, the instructor may even ask me to paint a familiar photograph or to replicate an existing work to train me? What’s the difference? I guess the courts will decide.
@@CNC-Time-Lapse If you were in art class, you'll start with how to hold your pencil sketching objects. Then maybe human pose with a real model. In the class I attended, the school prepare a premade artwork that is done by the instructor and asked to copy it. Of course might be different for other school since the one I attended was taught by Japanese teacher. And they are very careful of legal sides. We were only taught to copy or follow the material that he prepared. And that is only at the near end of class only after real life sketching and theories.
This small detail, that something must be made by a human in order to have copyright, is so important and is being overlooked by so many people. This video should have way more views.
The question then becomes, at what point can you change AI generated materials to be considered human made enough to get a copyright?
To be honest, I think this could empower small businessess. It's like typography which only people with a lot of money could afford. But nowadays changing a typeface is built into word. With AI Art you don't need expensive marketing consultants to make an Instagram/Facebook ad.
And regarding to art, I think artists are the og influencers. You don't really follow the art you follow the personality behind it.
We shouldn't worry about AI art, bots and deep-fakes are the real danger!
Yep! Most of art is the personality and story behind it. Eventually the shine of the AI art that everyone is using will die down and people will actually have to figure out how to use it well.
I don’t necessarily see a problem with you using it for your videos because Photoshop has existed long before which gives you the ability and flexibility to doctor images, albeit, from a person utilizing software and a computer. Where it gets iffy is when it’s used to doctor falsified imagery/news. What’s more scary is the technology behind deep fakes where people’s faces can be put onto another’s body. It’s quite a rabbit hole, and I don’t see it being de tangled anytime soon. It’s a complex and interesting conundrum 🤔
Totally agree. Like the one article I shared with them using ai photos in Syria/Turkey to scam people into donating. That's why we created our own policy so you can always check and ask if we are using it.
@@FrameVoyager if only more people were transparent. It’s scary when people use it for sinister reasons.
Love this, my creative partner and I have been talking about how we could best use AI to make what we do more efficient/increase our output/ect. Basically what you’re taking about, regardless if we like it or not, we should probably figure out how can use it to our advantage.
Pretty much. It's coming and the sooner you figure out how it works, the faster you can integrate it. With that being said, you can still choose to use it as ethically as possible and support rights for photographers
I think AI art (or any AI generated content) is more or less like saying to someone, I want this (whatever), could you create it for me as long as I can modify it how I want? Its not yours in that you didn't create it, you just asked for it and modified it to your liking. It doesn't take away genuine skills of being able to create artwork without it, but I do think it possibly could take away creativity when technology can mostly do it for you. Sure you could argue there have been tools around for years like Photoshop that allows you do manipulate images, but that is the person doing it, with skill, not a computer generating an image that you don't have to do much to make it how you want.
I also think that AI tech should not be exploiting other people's work as a basis for its learning, unless they give permission first as that to me it is a bit like forgery (when it is AI art) or stealing. Its unethical and I can see more debate and possibly laws introduced if AI technology like this keeps advancing and becoming more widely used.
I guess it depends on what you do with it? I mean I can point and shoot with an ARRI camera and get good footage but if I have no clue how to process it or edit it into something good, does it even matter how it was shot in the first place? I think a similar philosophy could be applied to AI art. Eventually, the market will become saturated with it and only the people who understand how to take the initial image or prompts and create something else with it will end up with anything to show for it.
For me, it's a tool for a small piece of what I do and is not the cornerstone of the content I make.
But I totally agree with you on the ethical side of it. Stable Diffusion just got a class action lawsuit, so while it's going to be an interesting legal battle, I think we will start to see some rules and regulations for this being put in place. Because AI is the next wave of technology whether we want it or not, companies have invested too much into it.
Thanks for the great comment!
The side broheim has about *needing* to work in conjunction with the AI, because otherwise you'd become uncompetitive is not a very POG point of view, essentially. AI creators want you to forcefully embrace it as a consolation gift for having everyone's work being used.
I totally agree that AI datasets should have regulations and pay royalties for work they've used. I'm glad to see the class action lawsuit against stable diffusion as that should hopefully start setting some precedents for the industry going forward. Better now than way later down the line.
But I wouldn't say you would uncompetitive, but I think it's a useful tool to understand even if it's not ready to be fully utilized yet.
But I totally get the sentiment, a lot of people feel the way you do about it.
@@FrameVoyager The problem I see with any sort of royalty payout is that due to the large number of images involved, even if there is a large pool of money set aside, the splits could be so small that it doesn’t really do that much financially for the artists and perhaps undercuts their argument in the process. Additionally, if we’re dealing with a generated work, if hundreds or thousands of artists are pointing to the same parts of the work and claiming “it copied me”, that also undercuts their argument by reducing the distinctness of their work.
As you can probably guess, I’m fairly skeptical of the lawsuit against AI art generators. I’m not completely unsympathetic to artists, and believe that they deserve credit and compensation for more direct usage of their work, but when you try to apply that to a statistical average/median of a large number of similar elements (which is kind of what these models are, based on my technical understanding), the logistics and arguments don’t hold up terribly well to me. I wouldn’t call myself a great fan of how the data was collected for these models, but I think they have a viable argument from a legal and to some extent even moral standpoint. We’ll see how it plays out in court.
I have a strong point of view about all of this; as does everyone else. How on earth will we determine what's right.
I swear it's like every year there are more and more ethical dilemmas haha
“A pretence of Art to destroy Art."
-William Blake
I definitely didn't expect a David Holz quote in the beginning
😅😅😅
The "big corporations taking away our art" argument falls apart pretty quickly when you consider that in a few years, users will be able to train these things themselves. Right now training via back propagation is extremely expensive and tagging requires a lot of human input, but this could change with for example, using the Forward-Forward method on dedicated analog ML hardware or other improvements, this is a rapidly changing field. Once people get their models trained themselves with an AI scraping the web and tagging the photos then the moral distinction of you learning something vs your computer doing it will be non existent.
I don't think it is complicated at all. It is fairly simple people should not be exploited, there is no ethical basis for exploiting people. So to train AI algorithms on artists work without compensation is unethical and I think there is a legal case to argue that it is theft. If you prompt an AI and it uses my artwork to generate something for someone else, I should be compensated, particularly when these companies have monetized the process. You have to pay for DALLE and Midjourney so they are making money and compensating none of the artist in their vast reserve of art. It is unethical.
If I could input my own art, if I could do a drawing or film something and input it into the ai and say make the drawing photo real...that is totally different. If I render a building or object and I say take this render and place it in this picture of x and the AI automates the process that is totally a different thing. What is is doing now is in no way shape or form ethical. the artist in their banks should be paid and every time their art is used to generate a prompt they should be compensated. The computer should know which image or set of images it uses to generate your prompt and some monetary compensation should go to the artist.
Also artist should have the option to opt out of this, like any random company can come along and sweep up your images into its training model with no notice to you and again, no compensation. I just don't get the confusion about this idea of having a monetized product that is generating revenue with no compensation to the artist whose work you are using to generate that revenue. It is plainly exploitation and it is unethical. All you have to do with this model is share in the revenue, make it so the AI keeps track of what it is using to generate the prompt and pay those artist...then what you'll have is everyone rushing to upload their work to your training module, and signing on to partnerships with these companies.
Totally agree. And honestly probably where this is all headed. But I do think the legal question is going to be a very murky one. Though I do get the sense from some of the ai companies that they will want to make some sort of agreement. Probably good for short term viability and to clear these specific ethical questions
@@FrameVoyager The thing is that the explosion of AI art generator companies (using Stable Diffusion AI motor as their core) that it is happening today has just one reason behind, making money before the party is off. They already know that they gonna end having legal troubles in the future (well for Stable Diffusion is a present thing) so they need to squeeze the AI til the last drop before the law forced them paying for each copyrighted image their AI use or, worst, to only relay in a dataset wich only contains non-copyrighted pictures, making their generation of images a couple steps back to current technology. The thing is that only a few current pro artist (probably less than 20%) gonna end using that technology, at least for a time, because as it is currently developed can't offer much to a pro artist, unless the style of the artist is very in pair with what Stable Diffusion or Dall-e can already produce, that apart from photorealism, the AI models usually have a very narrow spectrum of styles even if they were trained and feed with the majority of the artist from this world, even my artwork is in in the database used with Stable Diffusion and I can assure you that the AI can't produce any image near to what I can produce, it is useless for me unless I use as a reference in the same way I use Google or Pinteres (If I use them, coz usually I not). At the end is pure late capitalism, same shit as cryptocurrency and NFTs, be in the party first, make as much money as you can, and run away before things end badly.
@@Akenoomokoto Totally agree with you on the stable diffusion thing. And they just announced a class action lawsuit against that dataset. Which I am totally for. This is the time to figure out the legality issues and make sure that people get paid fairly for their work being used. I think a lot of this will have to be worked out because AI tech is going to be here to stay.
It's not where it needs to be yet for most professional use cases, but in many regards it's still really in its infancy and every time a prompt is made it gets better and gains more information.
I think it's going to be the future in some way. We need regulations and copright laws for it for sure, but I think this is different than crypto of NFT's. But totally get your point
@@FrameVoyager well popular NFTs like "bored ape" were made with AI generators trained by their authors. NFTs make some how popular the use of this kind of tools to make profit from the and actually is precisely in NFTs (and some websites like Society 6) where the AI bros are trying to make money (there is hundreds of videos here in UA-cam explaining how yo monetize AI images using NFTs technology and some other website, the majority of them promise impossible amount of money...). I'm not saying that AI art generators are the same thing, I mean, I said it, but I don't mean exactly that, what I mean is that this kind of tools and technologies are under regulated, have the same spread pattern, have the same late capitalism ideology behind, and have the same end users, at least in its initial shape, people who do not care about ethics (because NFTs are... Well also not very ethic) and the main interest is make some short of profit from them. I'm already aware that in the case of AI art generator software the target has a wider spectrum, there is people that just want to use for fun (but I don't see they paying for it) and also there is plenty of content creators that could use it for their video thumbnails, their Channel content and stuff like that, but... That's just collateral targets, or at least that's what I'm seeing now a days.
I'm all for innovation, but if the images used for generating the AI models weren't licensed by the original artists, don't the AI companies need to pay the artists for corporate use of their work? Not in terms of 'per image created artists are owed money' but in terms of the original use of the image to train the AI?
That's why there is a class action lawsuit against stable diffusion currently. It's more than likely going to set that very precedent. But not every company is doing that. It's kind of the wild west right now and hopefully we start getting some sheriffs in town haha
I certainly hope it reaches a point of agreement for all sides. I imagine in a few years Bing images will be a very different space both in search and legal terms (because of course we've all moved to Bing with Chat GPT 😉)
Great coverage of this by the way, loved reliving the Gorrilla lawsuit 😂
@@astrobotnautics5291 Same! But gosh if Bing's AI chat isn't going to be the one that ends us all I don't know what is haha. That thing is scary.
And appreciate it!
Personally I would only be offended if my original imagery is being used to create profit somewhere thereafter. Artists should be then compensated the same way they do with stock imagery. My work is being used in reproduction for profit? I want royalties. It’s one thing to talk about art and it’s another thing to talk business. Blue Willow obviously believes it’s all for the best it’s a business too. Can I ask an AI to attempt to draw up code for a program that takes their IP and uses it for whatever I deem “good for humanity” ? I don’t doubt the possibilities but it seems like a lot is being done in grey areas with little regard on ethics
Exactly. That's why I'm really hoping the class action lawsuit being brought against stable diffusion helps to start setting some precedent and rules and regulations. It's great tech, but it's the wild west right now
@@FrameVoyager I can stand behind those words!
Errrrrrrrr Koala BEAR. They’re not bears. Stop calling them bears.
Who called them bears?
AI art sucks because they never asked artist consent for widespread usage of their data. Not a single penny of compensation, while the AI Art companies reap the profits from peoples hard work.
Its a despicable practice that will lead to countless of lost jobs and its ALREADY happening in many studios. Countless of audiences & workers are creeped out by un-ethical business practice......because you say "but you can use it to further your work"....i tell you this. if you are artist working in the creative industries that LOVE making what they make, love the creative workflow you have.
By saying that, you are essentially telling us to ditch the thing that we so desperately love & crave to do. By saying that you are UNDERVALUING the worth of what artists and fellow creatives do.
I don't WANT AI to improve my workflow. I dont NEED AI to do that. I can improve my workflow with my own brain power. I am PROUD of my hard work and journey, but AI threatens to wipe that out. Further to this, creatives working in the industry DESERVE legal protections and regulations against AI companies scraping data without compensating those who made the data. It should be Opt Out by default, and Opt In is optional.....but the current model is the exact opposite.
Also, of course the CEO will side with his AI business practice. OF COURSE he would, its making him HUGE profits. This is a biased take.
W