• La Dame d'Azur@lemmygrad.ml
    link
    fedilink
    English
    arrow-up
    3
    arrow-down
    2
    ·
    2 months ago

    Consider it differently. OP never used the word art in their comment (the one you’re replying to). Not everything is art just because it’s drawn. As a designer I was never under the impression my logos were art in any way. Yet they still existed and solved a problem for someone. Therefore, we can say something doesn’t need to be art to exist and this frees up a lot of effort being put into dead ends.

    This is fair, though considering that the image uses the text of a poem I think it exists in a nebulously defined area between “art” and “not art” - but at the very least I can agree that OP didn’t explicitly mention art. That was perhaps just me jumping the gun; this is a bit of a hot-button topic for me as a writer.

    Are ai images art ? That’s not even a question I ask myself when I generate an image. I look for a specific thing when generating - how it handles text, “picture inside a picture inside a picture” and other technical details like that. At no point do I ask myself if this is art or not because I don’t need to answer that question like a sphinx riddle for the image to be born. We are experimenting, trying out different things, making a final shot out of a long process and you could argue this is part of the artistic process too but again I don’t even need to do that to keep doing what I do.

    Okay, yeah, all fair arguments to make.

    Its like we’re in two different physical places entirely. The people that absolutely demand the answer “is this art” (i.e. the ‘anti-AI’ crowd) and the people that are simply not (the ones that use these models). I’m far from the only person not caring whether what I make is art. I doubt many people, when they open krita or gimp, consciously think “hell yes I’m going to be making art today!” they just do their thing. To me, this argument betrays people as never having used AI image gen and actually contending with what it is. Though I agree on platforms like gemini and chatgpt you can just make a half baked prompt and the LLM will clean it up for the image model, but what I mean is that if one wants to have good arguments against AI, they need to find something better than “but it’s not art”, because the only response to that is “okay, I can still generate 6000 images per day though.” It’s not an effective argument and it’s a complete mismatch to what people who use these technologies currently do with them. It doesn’t speak to us whatsoever, basically.

    My only issue here is that art isn’t necessarily something you consciously decide to do - but other than that I can understand your arguments here. They’re all fair and reasonable and I confess to having had a pretty strong bias already. I may have been making an argument appear where it didn’t to and I’ll own up to that mistake.

    In fact you could take this further. Do you really need a human to make art? I could, at this very moment, make a script that queries deepseek with a prompt “you’re an artist that [description that deepseek came up with of itself as an artist]. you make AI art, but since you can’t directly use the image platform, you instead send out these various parameters as a Json object: prompt, negative prompt, seed, noise algorithm, scheduler, image size (etc.). This data is then passed through an API to the actual image generator who will generate the image for you”

    I don’t think you need a human - specifically - to make art but I do think you need a human experience (or something equivalent to it) to make art. What makes “art” art is what its creator imbues into it; thoughts, feelings, perspective, history, understanding, etc. Art is a reflection of its creator. For lack of better words it has a ‘soul’ to it that an AI cannot replicate. What AI can do is imitate art; it can’t create it. There’s no deeper meaning behind what it makes. It churns out whatever you put in the prompt. The AI isn’t ‘saying’ anything through what it makes nor is it expressing itself because it has no self to express. Thus it isn’t an artist and what it makes isn’t art.

    And then through some either prompting or scripting make deepseek into a ‘living’ artist that sends these out periodically when it feels like.

    Would that be art? Would deepseek be an artist? What would that mean for art?

    This is not a thought experiment anymore, this is completely doable right now by anyone who knows some python (or even the LLM itself could write this code now). Do we necessarily have to reckon with this riddle though to enjoy life and illustrative works? Not really.

    I think we do, actually. If some of the attitudes in this thread are to go by there seems to be a sizeable amount of people who think art is being gatekept by actual artists and that only through AI can all people make art - an attitude I find reductionist and technophilic. There is also the troubling notion of just surrendering human creative works altogether to automated machines - the very thought of which I find horrifying. As if automating culture itself is something desirable. I would expect this kind of argument from an Objectivist that wants even more proles stuck in the workplace. As many critics of AI have noticed: our culture is being automated while our labor remains manual when it should be the other way around.

    AI has an important place in the continued technological development of human civilization - this I do not question - but for art, specifically, it has no place. Our culture is a reflection of us; it is part of who we are as a species. It reflects what we feel, what we think, and what we experience in life. It needs to remain a wholly human experience lest we risk losing our very identity as a species. I don’t think this should be controversial to anyone who isn’t one of those extreme transhumanists that unironically thinks humanity needs to ‘ascend’ or whatever.