one thing that "AIArtists" do not know, is the fact that it does not matter if the image was made by a person or by an AI, if it is plagiarism, and the responsibility will be on those who share it even knowing this, AI is just a machine, the responsibility falls on those who share or sell the plagiarism regardless of the way the image was made.
Great video, boss. You raised a number of issues that everyone should be prioritizing. These companies have released abysmally unrefined products. And they've been successful because a lot of prompt enthusiasts have failed to consider the ethics- either intentionally or unintentionally. Anyway, the legal issues have not even been addressed by the justice system; the landscape of the Art industry has not changed; the current standard of generative art is not the future. In the meantime, Artists need to start: - copyrighting their work no matter the cost - learning how to evaluate their work so they don't get swindled, underpaid or sign a detrimental contract - sharpen their business skills and stop being complacent. The complacency is what got us into this situation in the first place; if we weren't so broke, someone would have sued one or more of these companies by now.
It's a very tough subject that is going to create many dilemmas in the field of creatives and its industry. I wouldn't say copyright is the way to go, but something needs to be done for this tech to not just be careless towards the art community. Thankfully, with the release of Stable Diffusion 2, Emad (Stabilty.ai's CEO) set a precedent of making all open source and collaborating with Spawning, which is creating a creators focused opt-in opt-out system. We should continue push ethical approaches for this technology because the programmers won't do it themselves.
I have been obsessed with Midjourney for 5 weeks, and getting praised in certain FB art groups for my art. Yours is one of the most thought provoking videos I have seen this year, and makes me question what I have been doing. Thank you 🙏🏽
The problem of course with plagiarism existed long before AI, but of course this has supercharged it. On good hardware you could do 16 000 images in the time it would take the original artist to paint just one. Such a weird time to be living in.
@@fabianmosele2321 I wonder if much of the anger and drama is created by "AI artists" who are new to the artist community and doesn't know the (un)spoken rules of what is acceptable to do. Wish they (including me) would be more humble, knowing where the power of these new toys/tools are coming from, and we are in someone else's playground and has to play by their rules...if we want to play nice.
@@Yownas73 Definitely. It feels their taking the tech / data collection / open source approach to a very different field. It is kind of irresponsible to disrupt the illustration sector while totally ignoring the fears of the community.
The problem is that people upload their artwork to services where the terms include the ability for that service to freely use the uploaded material. In that way, the artist already give consent.
It's almost like the Web2 problems around data collection and sharing are creating the new problems for the next decade of the internet that we couldn't have imagined.
Thank you for bringing these topics up. I also checked you webpage and you did an amazing job with the information about the GAN papers available! I hope you can make some tutorials on how you merge 3d with AI, your art is super interesting and also a great example of how AI generative algorithms are mediated by human creativity ❤
Thanks Claudix! I'll take your proposal into consideration and might make a tutorial of how to use 3D together with AI ;) Since I'm working on a animation blend between the two right now, I might as well post a behind the scenes for it!
Waaay ahead of you. I am using my own works, placing them inside with an art prompt style I like, example, anime, then see what I get and work with it. Not using anyone else's name, just my own. No copyright broken. But yes, I do agree if someone's name is being used, then I 💯 think that the artist of that name should have copyright unless it has consent from the artist.
Very nice to hear that! That's definitely making the use of AI more ethical and I think that should be the way to go with this tool. One might say though, that even though you don't spawn any artist's name, the SD model is already trained on copyrighted work, so even if you only say 'anime' it will create something that it learned from various copyrighted work.
My biggest problems with ai 1: Gatekeepung access to using certain ais to make art 2 Not sharing how to make cool stuff with ai... 3 My dreams are with ai that I CAN MAKE PICTURES OF MYSELF WITH AI / Artpeices using my face and body with ai layer . Your transitions were awesome i wanna learn how to do something like that. And that human evolution video was one of the best things i’ve ever seen with AI. but tutorials are usually boring or just hard to find. People keep their techniques SECRET nowadays and i understand why but my mentality is SHARE and progress technology forward. If u think people will make better art than u if they learn ur techniques then u are actually trash. I wanna learn as much as possible. I rly like touch designer rn cause i can put cool trippy GENERATIVE filters on my videos and AI technology rn is either too hard / boring to use or too EXPENSIVE. Can u tell when a photo or video is AI cause right now i feel like I CAN and i usually know what AI they used based on the textures in the photo and the way it looks is distinct with a lot of AI.
You got some points. I feel you frustration of keeping art techniques secret because I feel it's always much more meaningful when someone shares their creative process. As how I did those transitions, I used Stable Diffusion Deforum notebook. Another relatively easy to use interface for such things is the Stable Diffusion webui by AUTOMATIC1111. Plus if you want to put yourself in such AI imagery, you can use Stable Diffusion Dreambooth fine tuning. These need some time though for you to learn how to use them. They might seem too difficult at first, but trust me, if you put some time in it it will get very easy very fast.
Automatic 1111 is a great tool and you can already do what you wish to using dreambooth or training your own model (text inversion) from within automatic1111. I have done it with my family photos and they love it. Hardly any work required in Photoshop after as well. Great stuff. That said you really should invest in a good GPU so you can work locally on your own machine.
Thanks! I couldn't find much content around this on TH-cam so I thought I should do it myself. The problems are going to get worse over time so we should be prepared.
Stable Diffusion 2.0 is available now! Completely open source also this time the clip model is completely open source so the mapping of the Clip and LAION sets are very exact! This it the best artist toolset we have ever been given too!!
Once you start to give reasons for censorship no matter how logical they are then you open the door to more and more fine details being added all the time until virtually everything is censored. My advise to artists is adapt. Do not restrict creativity in the name of creativity.
There is fear of hard censorship on these models, which could happen if too regulated, but tech people also can't just disrupt the artist field and tell everyone adapt yourself without considering their opinions.
@@fabianmosele2321 I understand. I am an artist myself. No point moaning at the old, it's time to make opportunities' with the new. Let's be honest here. It is not just artists jobs at risk from AI, but nearly all professional and office based jobs. Where do we draw the line. Is the answer to ban AI. I would say the cat is out of the bag and it is to late for that now but it could be done.
@@chrislloyd1734 Of course AI shouldn't be banned, but it's in our own interest discuss the repercussion this tech has on artists. Since it is still in its infancy, the way we approach it now will impact the way it will develop in the future. haveibeentrained.com referenced by Swann is a good example of something that empowers the community of artists, in which they can give their choice if they want to opt-in or opt-out of the training dataset of future image models. It's not about artists loosing their jobs, it's about creating tech that fairly treats other people's data.
The mention of 'harvesting human creativity' is not something AI nay-sayers should toss off without defining 'creativity'. Humans are good at 'what if' scenarios but that ability doesn't exist in a vacuum. It started somewhere with a tiny subset of our current knowledge base. Hey Gorg, what if we take that thing you killed, remove the fur and throw it on the fire. It's pretty easy to argue that was the root of most recipes on the internet. So human creativity is just change over time. We remember, we document, we mix and match. AI creativity is no different, it simply got a huge head start thanks to our pressing need to record everything we do. Human creativity is not sacrosanct. It doesn't even exist in a void. AI 'creativity' is just as valid.
I do have a issue with the word creativity though. It is such a generic concept we like to throw around, that conversation around “is AI art creative” stay very surface level. Can someone be more creative with these tools, allowing them to create much more than before? Maybe yea. Are the commercial uses of this tool going to thrive creativity? I’d say probably not. In my opinion the “harvesting human creativity” mention is only meaningful the moment it gets used in the right specific context.
AI generated images will kill art. The fewer jobs for artist the fewer artist can afford to develop their talent. The fewer developed artist the weaker the community will be. This has already happened with the death of the Illustrator.
Imo I wouldn’t say kill art. Artists will always be demanded. It will change the dynamics and practices behind image creations though. Film making didn’t die cause everyone had easy access to a camera, but it did change the formats and content (like TikToks or TH-camrs).
@@fabianmosele2321 Everyone having cameras increased the pool of up and coming film makers. AI art will decrease the pool of human artists. Kill was hyperbolic so lets say severely cripple.
I remember thinking the same when software like Photoshop appeared. In the late 70's -80's it was all felt tip pens and paint brushes. Those were the days.
one thing that "AIArtists" do not know, is the fact that it does not matter if the image was made by a person or by an AI, if it is plagiarism, and the responsibility will be on those who share it even knowing this, AI is just a machine, the responsibility falls on those who share or sell the plagiarism regardless of the way the image was made.
Great video, boss. You raised a number of issues that everyone should be prioritizing. These companies have released abysmally unrefined products. And they've been successful because a lot of prompt enthusiasts have failed to consider the ethics- either intentionally or unintentionally. Anyway, the legal issues have not even been addressed by the justice system; the landscape of the Art industry has not changed; the current standard of generative art is not the future. In the meantime, Artists need to start:
- copyrighting their work no matter the cost
- learning how to evaluate their work so they don't get swindled, underpaid or sign a detrimental contract
- sharpen their business skills and stop being complacent.
The complacency is what got us into this situation in the first place; if we weren't so broke, someone would have sued one or more of these companies by now.
It's a very tough subject that is going to create many dilemmas in the field of creatives and its industry. I wouldn't say copyright is the way to go, but something needs to be done for this tech to not just be careless towards the art community. Thankfully, with the release of Stable Diffusion 2, Emad (Stabilty.ai's CEO) set a precedent of making all open source and collaborating with Spawning, which is creating a creators focused opt-in opt-out system. We should continue push ethical approaches for this technology because the programmers won't do it themselves.
I have been obsessed with Midjourney for 5 weeks, and getting praised in certain FB art groups for my art.
Yours is one of the most thought provoking videos I have seen this year, and makes me question what I have been doing.
Thank you 🙏🏽
I'm very happy to hear this video helped you questioning this tech. However I hope your journey in generative art won't be stopped by this.
Just be sure to be honest about your tools.
The problem of course with plagiarism existed long before AI, but of course this has supercharged it. On good hardware you could do 16 000 images in the time it would take the original artist to paint just one. Such a weird time to be living in.
True, many say this plagiarism already existed, but as you highlighted, this is on a whole different scale.
@@fabianmosele2321 I wonder if much of the anger and drama is created by "AI artists" who are new to the artist community and doesn't know the (un)spoken rules of what is acceptable to do. Wish they (including me) would be more humble, knowing where the power of these new toys/tools are coming from, and we are in someone else's playground and has to play by their rules...if we want to play nice.
@@Yownas73 Definitely. It feels their taking the tech / data collection / open source approach to a very different field. It is kind of irresponsible to disrupt the illustration sector while totally ignoring the fears of the community.
The problem is that people upload their artwork to services where the terms include the ability for that service to freely use the uploaded material. In that way, the artist already give consent.
It's almost like the Web2 problems around data collection and sharing are creating the new problems for the next decade of the internet that we couldn't have imagined.
Thank you for bringing these topics up. I also checked you webpage and you did an amazing job with the information about the GAN papers available! I hope you can make some tutorials on how you merge 3d with AI, your art is super interesting and also a great example of how AI generative algorithms are mediated by human creativity ❤
Thanks Claudix! I'll take your proposal into consideration and might make a tutorial of how to use 3D together with AI ;) Since I'm working on a animation blend between the two right now, I might as well post a behind the scenes for it!
Waaay ahead of you. I am using my own works, placing them inside with an art prompt style I like, example, anime, then see what I get and work with it. Not using anyone else's name, just my own. No copyright broken.
But yes, I do agree if someone's name is being used, then I 💯 think that the artist of that name should have copyright unless it has consent from the artist.
Very nice to hear that! That's definitely making the use of AI more ethical and I think that should be the way to go with this tool. One might say though, that even though you don't spawn any artist's name, the SD model is already trained on copyrighted work, so even if you only say 'anime' it will create something that it learned from various copyrighted work.
You should use stable diffusion and make a model with your own art. Would be cool.
@@_loss_ yeah,a model with volunteering and consenting artists would be amazing
My biggest problems with ai
1: Gatekeepung access to using certain ais to make art
2 Not sharing how to make cool stuff with ai...
3 My dreams are with ai that I CAN MAKE PICTURES OF MYSELF WITH AI / Artpeices using my face and body with ai layer . Your transitions were awesome i wanna learn how to do something like that. And that human evolution video was one of the best things i’ve ever seen with AI. but tutorials are usually boring or just hard to find.
People keep their techniques SECRET nowadays and i understand why but my mentality is SHARE and progress technology forward. If u think people will make better art than u if they learn ur techniques then u are actually trash.
I wanna learn as much as possible. I rly like touch designer rn cause i can put cool trippy GENERATIVE filters on my videos and AI technology rn is either too hard / boring to use or too EXPENSIVE.
Can u tell when a photo or video is AI cause right now i feel like I CAN and i usually know what AI they used based on the textures in the photo and the way it looks is distinct with a lot of AI.
You got some points. I feel you frustration of keeping art techniques secret because I feel it's always much more meaningful when someone shares their creative process. As how I did those transitions, I used Stable Diffusion Deforum notebook. Another relatively easy to use interface for such things is the Stable Diffusion webui by AUTOMATIC1111.
Plus if you want to put yourself in such AI imagery, you can use Stable Diffusion Dreambooth fine tuning. These need some time though for you to learn how to use them. They might seem too difficult at first, but trust me, if you put some time in it it will get very easy very fast.
Automatic 1111 is a great tool and you can already do what you wish to using dreambooth or training your own model (text inversion) from within automatic1111. I have done it with my family photos and they love it. Hardly any work required in Photoshop after as well. Great stuff. That said you really should invest in a good GPU so you can work locally on your own machine.
Great job on your video. Very good coverage.
awesome video on a very important topic and some very thought provoking takes aswell! ✨
Thanks! I couldn't find much content around this on TH-cam so I thought I should do it myself. The problems are going to get worse over time so we should be prepared.
Stable Diffusion 2.0 is available now! Completely open source also this time the clip model is completely open source so the mapping of the Clip and LAION sets are very exact! This it the best artist toolset we have ever been given too!!
Still not licensed for use
Also check haveibeentrained to opt in or out from the LAION database!
It's not available to everyone
Once you start to give reasons for censorship no matter how logical they are then you open the door to more and more fine details being added all the time until virtually everything is censored. My advise to artists is adapt. Do not restrict creativity in the name of creativity.
There is fear of hard censorship on these models, which could happen if too regulated, but tech people also can't just disrupt the artist field and tell everyone adapt yourself without considering their opinions.
Also check haveibeentrained to opt in or out from the LAION database!
@@fabianmosele2321 I understand. I am an artist myself. No point moaning at the old, it's time to make opportunities' with the new. Let's be honest here. It is not just artists jobs at risk from AI, but nearly all professional and office based jobs. Where do we draw the line. Is the answer to ban AI. I would say the cat is out of the bag and it is to late for that now but it could be done.
@@swannschilling474 Sorry... Not sure what you mean " haveibeentrained to opt in or out"?
@@chrislloyd1734 Of course AI shouldn't be banned, but it's in our own interest discuss the repercussion this tech has on artists. Since it is still in its infancy, the way we approach it now will impact the way it will develop in the future. haveibeentrained.com referenced by Swann is a good example of something that empowers the community of artists, in which they can give their choice if they want to opt-in or opt-out of the training dataset of future image models. It's not about artists loosing their jobs, it's about creating tech that fairly treats other people's data.
I WILL BE BETTER THAN ANY AI ART AND I WILL USE AI TO ACHIEVE THAT... IDGAF
I like ur spirit!
The mention of 'harvesting human creativity' is not something AI nay-sayers should toss off without defining 'creativity'. Humans are good at 'what if' scenarios but that ability doesn't exist in a vacuum. It started somewhere with a tiny subset of our current knowledge base.
Hey Gorg, what if we take that thing you killed, remove the fur and throw it on the fire.
It's pretty easy to argue that was the root of most recipes on the internet. So human creativity is just change over time. We remember, we document, we mix and match. AI creativity is no different, it simply got a huge head start thanks to our pressing need to record everything we do.
Human creativity is not sacrosanct. It doesn't even exist in a void. AI 'creativity' is just as valid.
License or shut up
I do have a issue with the word creativity though. It is such a generic concept we like to throw around, that conversation around “is AI art creative” stay very surface level. Can someone be more creative with these tools, allowing them to create much more than before? Maybe yea. Are the commercial uses of this tool going to thrive creativity? I’d say probably not. In my opinion the “harvesting human creativity” mention is only meaningful the moment it gets used in the right specific context.
AI generated images will kill art. The fewer jobs for artist the fewer artist can afford to develop their talent. The fewer developed artist the weaker the community will be. This has already happened with the death of the Illustrator.
Imo I wouldn’t say kill art. Artists will always be demanded. It will change the dynamics and practices behind image creations though. Film making didn’t die cause everyone had easy access to a camera, but it did change the formats and content (like TikToks or TH-camrs).
@@fabianmosele2321 Everyone having cameras increased the pool of up and coming film makers. AI art will decrease the pool of human artists. Kill was hyperbolic so lets say severely cripple.
I remember thinking the same when software like Photoshop appeared. In the late 70's -80's it was all felt tip pens and paint brushes. Those were the days.