I suppose it depends on how the person uses AI. If I have a specific vision and use AI to make that vision something I can share, why is it not my experience or intent? If I think it’s meaningful, for instance, to have an image of a glass vase full of bullets, does it matter if I drew it with a pencil, used a digital tablet, created it with a collage, filled a vase and took a photo or used AI image gen to create an image? Are all of them valid expressions of my idea except the last? And what if I personally feel that the AI image satisfies my intent and vision? Why is it nothing other than “algorithmically generated noise”? I’ve made lots of art in my time (was my major in college even) and I personally don’t believe there’s anything prohibiting AI from being a reflection of my intent and ideas. I also would say that anything hand made is only a reflection of my intent to the extent of my talent and capabilities at that time. In other words, you’re not seeing some unfiltered view into my soul, you’re seeing how well I was able to cobble together what was in my head which may or may not accurately reflect my view or experience or whatever.
But, that aside, lots and lots of art is basically just aesthetically pleasing or proficient at illustration. My bed sheet has flowers printed on it. An artist painted those flowers (at least originally, obviously not onto my sheets). I would call them an example of art. Not deeply meaningful and culturally relevant art but they ARE art in my estimation. But their only function is to make my bed look nicer. If I’m playing a game of Dungeons & Dragons and want an image of my wizard, I don’t need it to be carrying the weight and journey of my life or anything, I just need a visually acceptable wizard who looks like my character so I can say “Neat, here’s what my dude looks like”. So why not use AI image gen for that if it’s going to meet my needs? I would say that the vast majority of visual art we see or interact with daily falls under “there to be aesthetically pleasing or proficiently illustrative.” So even if someone thinks you can’t make Capital-A-Art with it, it still has tons of function.
To me that is decoration not art. Art is more than something that is pretty.
Again, IMHO art is an interactive communication between minds. Even if we can’t clearly articulate what we mean to be communicating and if the communication is received other than how it was intended.
To the degree that AI is being used as a tool by artists it is just another medium. And in all media there is more crap than great art.
My WAG is that the ease of making technically proficient crap with AI means even more crap with little effort and no feeling or thought behind it.
And the more it is just following a simplistic prompt the less there is a real mind on the creation side, the less it really is art, even if it looks the same.
Which I disagree with but, as I said, even if you want to say it’s “decoration, not art” then the answer to what the point of AI art is for would be “making decorations”.
AI is the wet dream of every no-talent hack that approached an established writer and said, “I’ve got some great ideas! If you write a book using them we can split the profits 50/50!”.
I have to say that although I don’t think I can add substantively to this thread, I have benefitted from it very much. It has really helped me understand art better. Mostly in how to separate the idea of art as an intentional creation versus purely aesthetic appreciation. I also feel like AI is generating the same discussion that photography did in the 19th century. I feel that AI art will always be with us. And like photography, there will be artists who wield it skillfully and hacks who pour out different levels of crap.
I don’t. I see art sort of like life. You have a bajillion amoebas and tube worms and not many pandas and bonobos but they’re all “Life” even if they’re not equally rare, complex or generally compelling. Likewise, doodles and painted flowers and sponge-prints on a shower curtain can be art even if they’re not as culturally important and compelling as Guernica or the Mona Lisa. Acknowledging a slime mold as life doesn’t devalue raccoons and calling the intentional decorative glaze designs on a ceramic teacup “art” doesn’t devalue The Pietà. It also avoids shifting boundaries of “Art is expression between people” “Ok, here’s a vase full of bullets I think expresses my thoughts on school shootings” “No, not like that…”
I’ve been messing around with some AI tools and I think that’s what they are. Tools, much in the same way a paintbrush, airbrush, or Autodesk Maya is a tool. And like any tool, there is a degree of craft that goes into using it. People dismiss “oh that’s just CGI” but to create a Marvel or Avatar movie requires mastery of dozens of different tools as well as a deep understanding of lighting, composition, textures, physics, motion capture and who knows what else.
The problem with CGI (and by extension GenAI) is that it’s easy to create bad art if you don’t know what you’re doing or take shortcuts. The audience might not know why it looks bad, but they can see it’s “off” somehow.
That’s the problem with a lot of GenAI content. It just looks off in the Uncanny Valley because AFAIK AI tools don’t really take into account physics, perspective, lighting, etc. Unless you maybe break in down into different components.
Or you know, it lacks any knowledge about how a device it’s depicting actually works. I saw an AI generated animation segment on a recent training vid that had a train that ran on overhead cables depicted. It rendered the train and the overhead assembly just fine, but neglected to render any of the overhead lines or its supporting poles.
That’s really just inattentiveness, ignorance or laziness on the part of the person generating the images (or video) though. If you get bad looking junk and say “Sure, that’s fine; let’s run with it” it’s not really an AI issue. Even if the AI isn’t correct about how a train looks, the person working on the video should know and should be continuing to work at it until it looks correct. Especially with modern services where you can take the animation with the missing cables and poles and say “This, but add cable and poles”.
True, but just about any human you asked to generate a train that worked via overhead wires would have known to generate the wires, if nothing else due to the fact that they would be prominent in any reference pic they worked from. I expect the person prompting this AI just asked for a commuter train, and didn’t know that there were different methods of sending power to them. It ends up being a human failure, but probably one that would not have happened without AI.
I think we generally agree with one another just with different perspectives. Someone inexperienced with with AI generation may assume the AI will just handle it especially if they view the AI as an entity rather than a tool. After years of playing around with it, I understand that choosing to use AI means that you’re shouldering the burden of handling it versus giving it to someone else.
The flip side of that is someone who actually knows what they are doing can actually put together an almost cinematic quality production more or less by themselves at a fraction of what it would cost to actually film it.
I came across something the other day that I thought was relevant to this thread. Sometimes I want to see a photograph (or whatever form of media) of how something actually IS. Not how it’s interpreted it should look by some algorithm. There is something inherently unsettling and untrustworthy about people who can’t understand the difference.
Is that documentary or artistic or what? It’s easy to make an argument that, between choices made at the time of shooting (staging, settings, lighting, etc) and post-processing (much of which is algorithmic these days) very little of what we see is honestly real. Heck, a large part of photography as an art is punching up reality in some fashion. And anything hitting a digital camera sensor is going through an algorithm, even in RAW.
But, if I want to see images of Native American arrowheads then I want to see actual photos of actual arrowheads and not Nano Banana’s trained concept of what arrowheads look like. And it would be silly to “Well, actually, those arrowheads were shot under staged lighting and adjusted in Photoshop for contrast…” to imply that an AI image is the same thing in that context. To say less of posting to a “Cool Places on Earth” thread with an AI images of a crystal-filled cavern behind a waterfall or nature-reclaimed power station. “Sure, my waterfall isn’t real but your iPhone internally adjusted the colors of your Alaskan cruise aurora borealis photos” doesn’t sound like a winning argument.
That said, my phone’s algorithmically adjusted photos of the northern lights are WAY more compelling than what I saw with the naked eye that night and I feel obligated to mention this when someone sees them.
I think it depends on the purpose of the image/video and how transparent the person capturing and displaying it is with respect to the media’s authenticity and intent.