The problem as i see it is that this particular drawing program doesn't
demonstrate depth and insight. So, sure, it sketches, but art is not about
capturing images: it is about providing insight, new interpretations, and
depth. Thus, a good photograph can be art, not because it captures the
image, but because the photographer a artist has exercised great care in the
selection, composition, and lighting of the subject matter, often revealing
something very different than would be experienced by a viewer of the
original scene.
Painting. My favorite artificial painting is by my friend harold Cohen.
Harold is a world-class painter (paintings in the Tate, for example), who
developed a computer program to paint. harold spent decades perfecting the
program, teaching it all he knew about art, developing complex rules and
interpretations.
Then he let it lose upon the canvass, and the results often surprised him.
The results were considered god enough to have allowed him to do exhibitions
(sometimes including real-time painting by the computer) in major art
galleries across the world as well as to get detailed art criticism in the
major art journals.
The new crowd of AI researchers seem to have forgotten him, perhaps because
he was not a computer scientist or AI researcher: he was, primarily an
artist. (Although he gave lots of talks at AI conferences).
The name of the program is Aaron. Harold says they are Aaron's paintings.
harold wrote the program, but he insists that the results are due to the
program, not to him.
see
bio:
http://www.kurzweilcyberart.com/aaron/hi_cohenbio.html
Aaron's latest paintings:
http://www.artnet.com/Galleries/Artists_detail.asp?G=&gid=566&which=&aid=605810&ViewArtistBy=online&rta=http://www.artnet.com
some earlier paintings
http://www.viewingspace.com/genetics_culture/pages_genetics_culture/gc_w05/cohen_h.htm
Nick Lambert's critical evaluation of Aaron as part of his PhD thesis at
oxford.
http://thesis.lambertsblog.co.uk/?page_id=237
Don
|