@patricialamas I've recently heard a talk from a programmer/designer who theorizes that AI art might actually get worse from here instead of better - because bad AI art is being harvested and put back into the algorithm, giving it bad data!
AI companies could develop a system to be able to recognize AI art in order to stop it from joining the dataset, but if they create this technology it would be a huge self-own, because other people could then use it to point out which images in the wild have been made by AI.
On top of that, the free Glaze tool by the university of Chicago was JUST completed and released, and allows us to "cloak" our images with invisible, tiny pixel-level variations that makes the algorithm read our style all wrong. If they feed your glazed images into an AI to replicate your style, they'll be getting Van Gogh like swirls instead 🙂
Glaze: https://glaze.cs.uchicago.edu/
AI bros are already reporting (complaining really) that it works as intended and that "the least we could do" is watermark our images to indicate they have been glazed because now, these images are poisoning the data and making it misidentify images. Boo hoo!
This reminds me of someone stealing their roommate's food then complaining that it's been laced with laxatives.
c55edfff-5dc9-4fb4-9994-958543dcfa4d-image.png
Anyway, the future of artists is looking bright!
