Latent space and GAN networks can be used as graphic and cultural research tools. In this case, a GAN has been trained on one of the most iconic images in history and subject to multiple variations: Leonardo da Vinci’s Gioconda. To create the dataset, digital copies of the original table have been used, but also all kinds of variations and versions, as well as the images of the public in the Louvre museum that appear when searching for La Gioconda in an Internet search engine. The result is a kind of synthesis of the imaginary around this painting and its fame in the world of today’s mass culture (searching for images by words in internet search engines has been the criterion for incorporating images in the dataset). At the same time, hypothetically, the network could generate infinite variations of the Gioconda, fulfilling the function and rhythm of mass production and consumption: the need for continuous novelties that are also always the same.
Experiment realized in relation to the exhibition A certain darkness curated by Alexandra Laudo at Caixafòrum.