AI art is challenging the boundaries of curation

Over the past few years, the number of artworks produced by self-proclaimed Artificial Intelligence artists has increased. Some of these works have been sold by large auction houses for very high prices. Initially spearheaded by a few technologically knowledgeable artists who adopted computer programming as part of their creative process, artificial intelligence has recently been embraced by the mass as image generation technology has become both more effective and easier to use without coding skills.

Technical progress in computer vision, a research area dedicated to designing algorithms that can process meaningful visual information, rides on the coattails of artificial intelligence. generative models are a subclass of computer vision software. Generative models are artificial neural networks that can be trained on a large amount of images. They can produce new images that are not contained in the original dataset after training. The images that were produced through this approach were somewhat lacking in detail, but they had a charm that caught the eye of many serious artists. A new model called DALLE 2 was unveiled earlier this year by the tech company Openai. DALLE 2 can imitate famous artists if the desired effect is specified in the prompt. There is a tool for the public to use for free.

The coming-of-age 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 made by AI—are not particularly original. These questions echo similar worries once raised by the invention of photography. By merely pressing a button on a camera, someone without painting skills could suddenly capture a realistic depiction of a scene. Today, a person can press a virtual button to run a generative model and produce images of virtually any scene in any style. But cameras and algorithms do not make art. People do. AI art is art, made by human artists who use algorithms as yet another tool in their creative arsenal. While both technologies have lowered the barrier to entry for artistic creation—which calls for celebration rather than concern—one should not underestimate the amount of skill, talent, and intentionality involved in making interesting artworks. Advertisement

The process of art-making can be changed by generative models. Artificial intelligence art blurs the line between creation and creation-related activities.

There are at least three ways in which art can be made with artificial intelligence. The first one has to do with the curation of outputs. Not all of the images produced by a generative program will be considered artistic. The process of creating outputs is very familiar to photographers, who often capture hundreds or thousands of shots from which a few, if any, may be carefully selected for display. The artistic process is different for photographers and artificial intelligence artists than for painters and sculptors. The act of cherry-picking good outputs is seen as bad scientific practice in the field of artificial intelligence. The name of the game when it comes to artificial intelligence is cherry-picking. In order to promote specific outputs to the status of artworks, the artist's intentions may be expressed.

Curation may happen before any images are created. Curation in art refers to the process of selecting existing work for display, while the work that goes into crafting a dataset on which to train an artificial neural network is referred to as curation. The network will fail to learn how to represent desired features if a dataset is poorly designed. The network will amplify bias if a dataset is biased. Garbage in, garbage out. The adage holds true for both artificial intelligence and art.

Advertisement

German artist Mario Kinglemann, one of the pioneers of artificial intelligence, has created a collection of thousands of portraits from the 17th to 19th century. He used this dataset to train generative algorithms that could produce an infinite stream of novel portraits that were displayed in real time on two screens. This is an example of an artificial intelligence artwork. The training data was a fundamental part of its conception. The artist's personal aesthetic preferences and taste are reflected in the final artwork, albeit through the distorted lens of the computer-driven generative process.

The ability to describe the desired result in natural language is a novelty spurred by the recent progress of generative Algorithms. This is known as "prompting" or guiding the program with text instead of random outputs. The illustration accompanying the article features several images generated by prompting DALLE 2 with the phrases "an artificial intelligence image generation system, conceptual art," and "an artistcurated artworks produced."