Is AI Art Coming To A Museum Near You?
Step aside Da Vinci.
Christie’s art house auctioned off a piece of pure AI art, valued at $432,500.
This is a substantial paycheck for a machine.
AI art is created through machine learning.
AI artists write algorithms learning specific aesthetics after analyzing thousands of images across the web.
Images from famous painters and artists including Da Vinci, Picasso, and Warhol.
An AI can combine the talents of all these artists through a few lines of code.
The History of AI Art
One of the earliest practitioners of AI art is Harold Cohen, who wrote a program called Aaron in 1973.
Aaron produced drawings following a set of Cohen’s rules.
After studying and learning Cohen’s rules, Aaron is now a free artist creating its own paintings.
The class of algorithms Cohen and others use are known as generative adversarial networks (GAN).
These networks run on a two factor system.
For example, a GAN will analyze famous paintings and produce similar renders, while discriminating against paintings that break rules.
The artist behind the machine is heavily involved in sifting through produced images, sometimes tweaking algorithms in response.
Completely Autonomous AI Art
AICAN or Artificial Intelligence Creative Adversarial Network, is a program that is almost entirely autonomous.
It can learn styles, aesthetics, and create art of its own.
One of AICAN’s pieces sold for $16,000.
AICAN even understands the value of novelty in art; never copying entire styles while refraining from being too novel, which may turn off viewers.
This makes AICAN’s art unique and acceptable.
A.I Art Beyond The Canvas
And A.I artistry doesn’t stop at the canvas.
Google’s latest art project “Poem Portraits” takes a word of your suggestion and generates a unique poem.
So if you need a little help penning a few words for your significant other on Valentine’s Day, Google can be your co-author.
The algorithm is trained via deep learning neural networks on 20 million words of 19th Century poetry.
The AI doesn’t copy or rework existing phrases, but uses words it studies to create a complex statistical model.
The poetry can either be profound or completely meaningless depending on the word you choose.
AI is also taking over producer booths.
Meet AIVA, which composes soundtracks for movies, commercials, games, and trailers.
The scores it produces are indistinguishable from today’s human composers.
Currently, the algorithm is an opportunity for filmmakers and producers who want control over their own score.
AIVA comes loaded with preset algorithms to create its own music.
Maybe someday soon we’ll be watching an entire YouTube video, or reading a thoughtful philosophical article from an AI author.
Hopefully, not too soon.
