Raw LLM Responses

I think the antagonistic nature doesn't do anyone any favors on either side. There are pros and cons like anything. First to cast a wide net and say that AI and the use of it is theft is disingenuous. It in and of itself determined by the images that are used in the data set. Just for one example? The data set in stable diffusion (The LAION Dataset) consists of something like 2.3 billion images. The only way someone would get an image that is even remotely similar to an existing image is if they were to feed the image to the AI via img2img and prompt it in such a way. Shad talks about this in his one video and on this I feel he is quite fair and articulates that theft is a problem with the end user. Not the tool or technology itself. Nothing is store in that data set save for the recognized patterns from the diffusion process. I went to art school. Got a degree in graphic design. And guess how we learned to draw, learned to design logos? We studied the works of other artists. In Drawing 2 we had to create a "Master copy" from an existing work in charcoal. The "learning" in that respect is not unlike how we as humans learn to draw by looking at the styles of other artists. I've seen some people in the comments below say they were learning a particular style and that's fine. There are very few "Original" ideas. Know who else said that? “Good Artists Copy; Great Artists Steal“– Pablo Picasso "There is a story about Igor Stravinsky (1882–1971), the Great Russian composer and pianist. He was about to start a new ballet. Instead of starting from scratch, he decided to pull up with some of his favorite classical composers and started correcting their baselines and melodies as if it were his own. In the end, he came up with his own set of harmonies and compositions masked from those exposed great works. He indeed quoted, “Lesser artists borrow; great artists steal”. Many people confuse copying with stealing, Copying is the exact duplication of someone else’s work, on the other hand, stealing is nothing but iterating those nuggets of ideas that already exist to build that up into something new which never existed." That isn't to condone the act of stealing or trying to pass off another's work as your own. But think of it a bit like fair use on Youtube. Is it transformative? Sticking with Picasso. Remember Starry Night? How many derivative works have been done on Starry Night? I don't feel that AI art or traditional art need to be at odds with one another. I do think there are moral and ethical concerns to be sure. But is it "theft" if someone creates an image in the style of another artist? No. Because the image that is produced is a derivative that is created from the recognized patterns in the diffusion process. Again the only way an image could be considered to be "Stolen" is if the end user feeds an image directly via img2img. I think that AI is a new medium and another tool that even traditional artists can implement into their toolkit. Let's say you have a particular style all your own. You could then train the ai on your own style and further increase your workflow in producing conceptual imagery, etc. I can understand the concerns regarding whether AI will replace traditional artists. But this is a problem with corporations in general. Hell I worked in a factory and got replaced by a machine. So I understand that. The thing is though that just because AI exists doesn't mean people will stop or should stop producing art if they enjoy it. Am I worried about AI taking my job as a graphic designer? Not really. In fact with Adobe? AI has aided my work. What used to take me a 15-20 minutes to clone out someone from a photo? I could do with generative fill. I still had to select the area desired. There was a lot of handwringing and worry about photography too. And last I checked traditional art survived photography. Photography even allowed for new ways for people to be creative. I also think of the implications for people who may be physically unable to create art traditionally. Take blind people for example. Can a blind person now create and express themselves through AI using speech to text? Just as another example? Sean Aaberg was an illustrator who had a successful kickstarter for "Dungeon Degenerates" A tabletop board game. In sept. 2018 he suffered a stroke that affected the left side of his body (His drawing side). He took to using Midjourney as a creative outlet. 61 million people in America have a disability. That's 1 in 4. They noted there is no doubt about craftsmanship or the human touch, but that AI has helped them achieve creative feeling they hadn't had for 4 years. Lastly, we have to ask ourselves "What is art?" This is a question that has been asked before. Medium has an article written by Karistina Lafae called "Gatekeeping in the Art World and AI tools as accessibility." She asks several questions... Is a picture of a can of soup art? Do the things your child makes in art class count as art? If not, are you going to tell them that? How would you even explain it? Because it’s not “good” enough? Because they haven’t spent years honing their craft? Because it’s a mess of construction paper and scissors and glue? Are superhero comics art? And if not, why are comic artists hired to draw them? What about fan art? Do the thousands of fan artists who make commissions for other fans not count as artists? Do they not put their own imagination, skills, style, and soul into their creations, even though they don’t have permission from the people who own the copyright to those characters? What about Marcel Duchamp's The Fountain? I studied Art History as part of my major in college. And the thing about Duchamp and the Dada movement was that for the longest time, their work was not considered art. Duchamp at the time challenged the preconceptions of what we perceive art to be. In many ways, Duchamp was the avant-garde. He is considered one of the most influential artists of the 20th century. You may also remember him for the mustache on the Mona Lisa. Duchamp was more concerned with the idea rather than the process. In that respect? It still takes someone to prompt the AI with an "Idea".