If you’re a starving artist who at least finds consolation in the fact that a computer will never take your job, you may need to think again.
As reported by the Wall Street Journal, last weekend Google held an art auction in San Francisco for works created by its algorithms. The most expensive image sold for $8,000, and all proceeds from the auction and event went to the Gray Area Foundation for the Arts. Examples included pieces that imitated styles of artists like Vincent van Gogh, and each image was printed on paper.
Eleven artists, including some who are engineers, created the auctioned art using variations of Google’s DeepDream code, which the company open-sourced last summer. The artistic process uses the same neural network idea as that which was used by Google Photos to identify images in its search function. But instead of asking a network what it sees in a photo, it is asked to see something in a blank image. (Google’s Research Blog goes into great detail on this process.)
This is hardly the first time algorithms and artwork have collided. Examples of computer-generated algorithmic art go back to the 1960s. At last summer’s International Joint Conference on Artificial Intelligence, one theme was the relationship between AI and art, and an exhibition of AI-made art ran alongside the conference.