Raw LLM Responses

Like the youtuber analogy, it depends on the piece. Sometimes a good looking piece is all it needs to be, and I don't care about the creative process behind it. In that case, AI art is fine to me. Other times the AI art is just super bland, which is pretty common, but also not something that is inseparable from things that humans make. It isnt any more bland because it was made by a machine. I don't have the technical artist mindset to find intrigue behind brush strokes and why they were placed a certain way. I don't usually ask why a pixel was placed in a particular spot unless it bothers me, or is already a particularly compelling piece. I simply think "they are professionals and put the strokes in those spots because that is what would look the best" However, the best art I've experience always has some kind of craving behind it, and you can feel it through the piece. I have yet to see an AI piece with those levels of visceral experience behind them. To that degree, I tend to agree with you, AI art is boring. Even the most impressive technical pieces do not compare to the passionate expression of an ametuer. Though this is comparing ants to gods. Not all pieces need passion, and art can be appealing without that level of soul. Personally I think its not a limitation of AI as a whole, but a limitation of how AI generates things today. It has no "life experiences" to draw from, only training. Similar to how many people say you must experience life to be a good artist, AI needs to do the same. An artist who only creates based on training without exposing any of their soul will only produce mediocre works. They may be impressive, but they won't be interesting in the same way. AI art has the potential to be fascinating because it is the creation of something specifically not human, in the same way an elephant's painting or a parrot's songs are interesting. In my eyes, these types of non human art make the technical stuff interesting. That is when I ask "why would a non human thing put a stroke in that spot?". AI as it is now is just a best fit algorithm, so the answer comes pretty bland like you said. However, if Vedal were to give Neuro-sama the ability to draw, I would probably find those creations a bit more interesting. ...and neuro is still just a predictive text model.