Giving competition to creator’s competition

When the Mexican muralist Diego Rivera — whose full name is Diego María de la Concepción Juan Nepomuceno Estanislao de la Rivera y Barrientos Acosta y Rodríguez — was told by a critic his birds did not “look like flamingoes the way God made them”, the unfazed artist replied, “No — they look like the way I made them.” Underlying the assertion that nature follows art is the credo that the handiwork of the Creator can legitimately be recreated, with value addition, by a human agency as represented by the artist. Now it seems that the creations of artists can, in turn, be recreated by an agency other than human: AI, or artificial intelligence.

Using a method called generative adversarial network (GAN), which involves feeding thousands of images representing a particular genre of painting into a computer that then assembles a composite facsimile, AI artists have produced works that include a portrait in the classical European style that fetched $432,500 at a recent New York auction, a price tag 60 times higher than the initial bid. A similar technique has also been used to produce AI artworks that are displayed on screens and change constantly as new data is fed into the computer. While purists might dismiss such ‘paintings’ by saying that whatever they are, art they AIn’t, practitioners of GAN are laughing all the way to the Braque.