I compared AI generated images with my own artwork to see how AI performed. I used the text-to-image generation ability of DALL-E and DreamStudio Lite to enter in the intended themes of my art. I tried a variety of text descriptions, including the themes I intended to evoke, to guide the AI system to recreate my artwork.
I failed in this pursuit. The comparative images below show the closest matches. I do recognize that while the artist is trying to communicate ideas with his or her artwork; however, the viewer is part of the process. One mustn’t de-art the art. Perhaps this AI failure is a good thing for artists?
No comments:
Post a Comment