AI, Art and an Auction
- Last week, as Christie’s announced it would be the first auction house to sell art made by artificial intelligence, Richard Lloyd, of the firm said: “We are offering a public platform to exhibit an artwork that has entirely been realized by an algorithm.“
- The work is developed by French art collective Obvious using a technique called Generative Adversarial Networks. The art will be on sale October 23-25.
- Hugo Caselles-Dupré, a co-founder of Obvious, said “We really believe that AI can be a new tool for art, In 1850, when the camera showed up, it was only used by highly qualified engineers and so it was not considered for its artistic potential. We think we are in the same situation, because people view us as engineers, but we really think this type of technology will be used more and more in art.”
- Gauthier Vernier, another Obvious co-founder said in a blog post, “With AI, parts of the creative process are getting automated. And those parts are the least complex ones. New creative tasks are appearing, and creatives will soon have a new tool, an algorithmic pencil, brush, drum pad, allowing them to have a creative machine in their hands and gain in efficiency.”
- Regarding Obvious’ AI artwork: It is based on data sets from 15,000 portraits painted between the 14th and 20th centuries that were processed with AI algorithms to create the work. While some people may find the artwork blurry or too abstract, the use of AI in the arts will increase and techniques will become more refined. The future of art will include autonomously created AI works, as well as artists using creative tools that are AI driven.
- Regarding the value of Obvious' artwork: Christie’s suggests the art could sell for around $10,000. Similar to other areas of art, factors driving the price will include views on its current and future value, its condition, the brand of the artist and, of course, supply and demand. We’ll have to check back in late October to learn its purchase price.
- Regarding AI art and the law: Historically, copyright protection for art has applied to original works created by humans. With AI, questions about ownership of data sets, patents on algorithms and other factors will likely need to be considered.
- Regarding Generative Adversarial Networks (GANs): GANS is an AI approach introduced in 2014 by Ian Goodfellow (currently a Google). To date, the primary use of GANs have been in the area of image processing.