Réalité, possibilité, nouveauté : ce qui reste de l’inspiration artistique dans les images générées par l’Intelligence Artificielle. Esthétique et sémantique
This article focuses on images generated by Artificial Intelligence via services such as DALL-E, Midjourney or Stable Diffusion, which convert natural language into images. The article is not concerned with whether these images can be considered works of art in terms of aesthetic quality, authorship...
Uloženo v:
| Hlavní autor: | |
|---|---|
| Médium: | Article ou chapitre numérique |
| Jazyk: | Français |
| Vydáno: |
2024
|
| On-line přístup: | Accès Université d'Orléans et IFPM Accès Université d'Orléans et IFPM |
| Shrnutí: | This article focuses on images generated by Artificial Intelligence via services such as DALL-E, Midjourney or Stable Diffusion, which convert natural language into images. The article is not concerned with whether these images can be considered works of art in terms of aesthetic quality, authorship or artistic originality. It is concerned with whether generative Artificial Intelligence can create under the same conditions that were once prescribed for art: the ability to represent things not only as they are, but also (or especially) as they could and should be. However, the exploration of possibilities by art or Artificial Intelligence can be understood in many ways: creation according to the criterion of verisimilitude, the capacity to idealize reality or to broaden our ideas and notions, or the capacity to liberate the imagination in order to produce the fantastical or the surreal. Although Artificial Intelligence tends to reproduce the aesthetic or social stereotypes that prevail in databases, it can nevertheless demonstrate a certain capacity to produce novelty in the ontological (and not the aesthetic) sense of the word, by enabling us to visualize new instances of notions we cannot otherwise define, such as an artist’s style or vague concepts with fluid boundaries and probable properties. |
|---|