Only for a In a few years, the number of works of art created by artists with artificial intelligence has increased dramatically. Some of these works were sold by major auction houses at staggering prices and entered prestigious curatorial collections. Originally led by a few tech-savvy artists who adopted computer programming as part of their creative process, AI art has recently reached the masses as image-generating technology has become more efficient and easier to use without programming skills.
The art movement of artificial intelligence is based on the technical progress of computer vision, a field of research devoted to the development of algorithms that can process meaningful visual information. Central to this story is a subclass of computer vision algorithms called generative models. Generative models are artificial neural networks that can be “trained” on large data sets containing millions of images and learn to encode their statistically significant characteristics. Once trained, they can create entirely new images not contained in the original dataset, often guided by textual prompts that clearly describe the desired results. Until recently, images created using this approach remained somewhat lacking in coherence or detail, although they had an undeniable surrealist charm that attracted the attention of many serious artists. Earlier this year, however, technology company Open AI unveiled a new model called DALL·E 2 that can generate highly consistent and relevant images from virtually any text prompt. DALL·E 2 can even create images in certain styles and quite convincingly imitate famous artists, if the desired effect is adequately defined in the prompt. A similar tool was released for free under the name Craiyon (formerly “DALL·E mini”).
The rise of AI art raises a number of interesting questions, some of which, such as whether AI art is really art, and if so, to what extent it is really created by AI, are not particularly original. These questions echo similar concerns that once fueled the invention of photography. By simply pressing a button on the camera, someone with no drawing skills can suddenly capture a realistic image of the scene. Today, a person can press a virtual button to launch a generative model and create an image of almost any scene in any style. But cameras and algorithms don’t make art. People do. AI art is art created by human artists who use algorithms as another tool in their creative arsenal. While both technologies have lowered the barrier to artistic creation that calls for celebration rather than care, the level of skill, talent and dedication involved in creating interesting works of art should not be underestimated.
Like any new tool, generative models bring significant changes to the art-making process. In particular, artificial intelligence expands the multifaceted concept of curation and continues to blur the line between curation and creativity.
There are at least three ways in which AI-assisted art creation can involve curation. The first, and the least original, is related to the curation of results. Any generative algorithm can produce an indefinite number of images, but not all of them are usually given artistic status. The process of curating results is familiar to photographers, some of whom regularly take hundreds or thousands of images, of which few, if any, can be carefully selected for display. Unlike painters and sculptors, photographers and artists with artificial intelligence have to deal with a large number of (digital) objects, the curation of which is an integral part of the artistic process. In artificial intelligence research in general, the act of “selecting” particularly good results is seen as bad scientific practice, a way of deceptively overstating the perceived performance of a model. However, when it comes to the art of artificial intelligence, cherry picking can be the name of the game. An artist’s intentions and artistic sensibility can be expressed in the very act of promoting certain results to the status of works of art.
Second, curation can also occur before any images are created. In fact, while “curation” as applied to art generally refers to the process of selecting existing works for display, curation in AI research colloquially refers to the work that goes into creating a dataset for training an artificial neural network. This work is critical because if the dataset is poorly designed, the network will often not learn to represent the desired features and perform adequately. Furthermore, if a data set is biased, the network will tend to reproduce or even reinforce such bias, including, for example, harmful stereotypes. As they say, “garbage in, garbage out.” This adage also applies to AI art, except that “garbage” takes on an aesthetic (and subjective) dimension.