Hello! This will be the first of our weekly discussions. This thread will be pinned in the local timeline all week long and I hope that everyone feels encouraged to voice their opinion.

This thread will be locked after a week. Make sure to join the discussion in time!

This weeks topic is: AI “art” and its impact on the furry fandom and artists.

Here are some guideline questions that you can use:

  • What are your opinions on AI generated text and images?
  • Should furry spaces allow AI generated images?
  • What do you think are the use cases for these tools?
  • What long-term impact do you believe it will have on the fandom and the fandom’s artists?
  • Wander@yiffit.netOPM
    link
    fedilink
    English
    arrow-up
    6
    ·
    1 year ago

    For me AI generated images are double-edged sword. As a non-artist I find that these are an amazing tool that allows me to create the specific type of Yiff that I really enjoy. However, I can definitely empathize with the concerns voiced by artists.

    To be honest, I have to say that I don’t believe that learning or training from publicly available artwork is immoral, especially if it’s not for profit, I am worried that many of the popular models seem to be clearly ‘over-fitted’. Over-fitting happens when you train a model so much that it ends up reproducing content that seems almost copy/pasted instead of merely being able to reproduce certain concepts present in the prompt.

    The second concern I have is related to mediocrity and saturation of AI imagery. I believe that in upcoming years many of us will grow tired of seeing AI imagery as a replacement of stock images. The truth is that most images generated aren’t really that good and people will quickly grow tired of seeing them pop up everywhere. It will become a sign of mediocrity or poor quality.

    However, I also think there’s a silver lining: AI images have the chance to drastically increase the amount of visual support that any type of content has. This could lead to us growing into a society in which images and other forms of visual representation such as mascots might become more and more popular (as it is in Japan, for example). And, as a consequence, this could increase the demand of quality artwork produced by real artists.

    In regards to the furry fandom, I think it’s a good idea to separate AI images from artist-drawn artwork. Not only because of the controversy, but also because of the potential for massive amount of low-quality an low-effort content.

    • Crazy Pony@rubber.social
      link
      fedilink
      arrow-up
      6
      ·
      1 year ago

      @Wander

      Posting this from beeing not an artist

      I do see potential to use these image generators to quickly try out concepts and scribbles when i have no idea how something should look, but try out how it might look

      Then later i could give it to the real artists as the refsheet