• mfed1122
    link
    fedilink
    English
    arrow-up
    3
    arrow-down
    6
    ·
    1 day ago

    Okay, yeah I totally agree that text translation and data interpretation is a more obviously acceptable use of AI. Definitely when it comes to art it gets muddier. I hate a lot of the ways AI art is used, treated, and understood, for what it’s worth.

    I do think, though, that while art is valuable as an expression of the artist’s emotions, it’s also fair to say that part of art’s value is the impression it makes on the viewer. If I see an artwork of a beautiful scene, it’s maybe less about this being an expression of the artist, and maybe more about how the scene makes me feel a certain nice way. In fact I think many artists actually strive to “remove themselves” from their work, they actively don’t want their work to be a reflection of their own emotions, but instead they want it to produce a certain effect in the viewer.

    Is Romeo and Juliet valuable because it tells us how Shakespeare thinks and feels? Or is it valuable because it makes us think and feel things? Of course there’s a mixture there, but I think it’s mostly the second one.

    If a bunch of rocks tumbled off a mountain and struck a xylophone and coincidentally played Pachelbel’s Canon, don’t I still hear beautiful music?

    So yes, I agree that while AI art is weaker in the dimension of enjoying thinking about the artist’s intent or the reflection of the artist’s mind, I think it can still fulfill the other great calling of art. And for course, it does reflect the prompter to some extent - albeit of course a much lesser one. But we also have human artists create aleatoric music and artworks (often specifically to efface themselves from the process), and we call this beautiful and creative, too - but arguably this is no different than leaving elements of the composition up to the chance of the generation’s random seed.