I see where you're coming from and I like your comparison to sampling, I guess I'd have less of a problem with it if people who's art is being used to generate these images got the contribution and money if it's being monetized, like with samples. Because ultimately the ai isn't creating anything out of thin air from what I understand. It's fucked up to me that people can create a "banksy" of their own based on his works, generated using images of his as a reference and he gets no credit or cash from it.
I‘m glad you introduced these things into the topic, because they are relevant. Everything you said is right. Your point about sampling is well taken. At some point, these companies…Midjourney, Dalle2, Nightcafe, etc need to face the wrath of the artists and be made to abide by some sort of regulation. Text-to-image models like Stable Diffusion are making it possible to not just spoof but to actually straight up replicate the works of artists- easily. Yesterday, I just entered a simple prompt of “man screaming” into Midjourney and in literally 2 seconds it gave me back a picture-perfect replica of Edvard Munch’s famous painting, “The Scream”. I wasn’t even looking for that specifically but it delivered it to me anyway. There is currently no art equivalent of sample clearance when it comes to AI generators and that’s not good. Even without AI, one artist copying the style of another’s work is (rightly) frowned upon. Ultimately, in a way it is a slap in the face to the processes and work of artists.
Also, Midjourney and these other companies are enriching themselves on the backs of artists by offering use of their product as premium subscription services.
i think it‘s not an unfair question to ask people who use it and monetize what they prompt for, “can you make an image like this
without using an AI generator?
I’m a Photoshop wizard. Been using it to create art and pro graphics for almost 25 years. There are some techniques on PS that I developed more than 20 years ago that I still have not seen others utilize. There’s another specific technique that I authored in 2005 and used to publicly make a certain kind of photoart that I shared on DeviantArt and other places. A couple pieces went kinda viral. 10 years later I began to see another guy who I know as a fact followed me figure out how I did some key aspects of my technique and started not only making stuff that was directly imitative of my art but began to pass it off as HIS innovation and unique inspiration. His expression of it was cheap, uninspired, hasty and not tasteful. It lacked soul, uniqueness and depth. I hated that shit. So I do understand the pain of fellow artists.
For these reasons, even though the tech is exciting, at the same time I feel conflicted when it comes to AI art generation. On the positive side, I’m able to use it to enhance what I already do by utilizing it as an artistic tool along with the others at my disposal. It can make my workflow easier in some regards. I still believe that for the most part individual image results are drawn up from perhaps thousands of images. No one is a complete original with no outside influences. Looking at Michael Jackson one could see Frankie Lymon, Jackie Wilson and James Brown. Looking at Prince, over his career one can see Jimi Hendrix (long coat with ruffled collar shirts, structure of his guitar solos), James Brown (dance moves, mic tricks and grooves) and Little Richard (facial expressions, hairstyles and gender ambiguity)- yet he took those influences and his own perspective and created his own aesthetic and style. Influences are influences. But even then, if Prince performed or recorded a JB song, JB or his estate would get paid. Sampling is more ambiguous as compensation goes. Artists go to court all the time for unauthorized usage of their songs, but at least there is the law stating “get a license or do not sample“ that original music artists and bands can stand on when suing for unauthorized usage. There is that framework in place.