The Sci-Fi Gene occasionally pauses to observe respectfully the evolutionary progress of our future overlords and masters as we approach the inevitable singularity. It seems that, while we were metaphorically sleeping, an Artificial Intelligence-generated image was entered into an art competition in Colorado and won first prize, to the annoyance of some human artists and commentators.
The picture, Théâtre D'opéra Spatial was produced by Jason Allen by sending an unspecified prompt to the art generation AI Midjourney. You can read more about the picture and the online reaction at VICE here.
Personally I'm not sure what the fuss is about. All the AI did was produce a detailed and evocative picture in classical style with powerful composition and lighting, layers of possible meaning and symbolism and an emotional kick. It's not like it taped a banana to a wall or anything.
As far as I can understand, AI artists work by analysing large numbers of images, breaking down the details in the images and deducing rules about the way those details go together, then using those rules to create new images. (Of course this is completely different from art students learning by studying great masterpieces.) A consequence of this process is that it seems to produce Impressionist works. For example here are two responses to my prompt "Renaissance woman with mysterious smile" by the AI Wonder and, for comparison, a little-known work by a human artist on a similar theme.
From a distance, the AI images look convincing enough. It's probably a good idea not to count the fingers too carefully, look too closely into their eyes, or speculate too deeply on exactly what lies behind their expressions. In all three cases perhaps it would be better not to know.
Posts
No comments:
Post a Comment